Gen Con?

So hey--for those of us too lazy to read the blogs: Anyone win yet? Peter D Bakija pd...@lightlink.com http://www.lightlink.com/pdb6 "So in conclusion, our business plan is to sell hot, easily spilled liquids to naked people." -Brittni Meil

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF2E730D.215D5%pd...@lightlink.com... > So hey--for those of us too lazy to read the blogs: > > Anyone win yet? > > > Peter from Michigan was the winner, with 1.5 VP in the final table. (I didn't play a game with him myself, so I don't know his last name.) He played Arika & friends. Jared Strait was #2, with 1.5 VP as well, playing Nos Princes. Ben Swainbank was #3 with 1 VP, playing G2-3 Gio with Le Dinh Tho, Gio allies, & media locations. 4 & 5 were Josh Duffin (classic law firm) and Stefan Ferenci (cel/CEL guns). I believe Josh was 4, but I don't know for certain, as I didn't see the initial seating of the final table. There were 63 people at the first round of the NAC; it took 0 GW, 2 VP, and some (unknown to me) number of TP to qualify for the top 40 on day 2. - Pat

Pat wrote: > Peter from Michigan was the winner, with 1.5 VP in the final table. (I > didn't play a game with him myself, so I don't know his last name.) He > played Arika & friends. Man. Did the national championship end with a time out? [ quoted text not captured ]

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF2E8C3B.215DD%pd...@lightlink.com... > Pat wrote: > >> Peter from Michigan was the winner, with 1.5 VP in the final table. (I >> didn't play a game with him myself, so I don't know his last name.) He >> played Arika & friends. > > Man. Did the national championship end with a time out? > It did... but not from lack of trying. Jared & Peter were going after each other up until the very last minute. I believe that Peter had the votes in hand to finish Jared off. He was in the process of calling something when time was called (KRC, maybe?), and had something else pretty nasty (Anarchist Uprising, maybe?) in hand, IIRC. I didn't see Jared's final hand, so I don't know if he had any 2nd Trad to stop a potentially ousting vote. - Pat

Pat wrote: > It did... but not from lack of trying. Jared & Peter were going after each > other up until the very last minute. That is good to hear. Like, full well realizing that having no time limit is a recipe for madness, but I'd really like, at the very least, the National Championship not end in a time out. But ya know, it is still a total pipe dream :-) [ quoted text not captured ]

Pat wrote: > "Peter D Bakija" <pd...@lightlink.com> wrote in message > news:BF2E730D.215D5%pd...@lightlink.com... > > So hey--for those of us too lazy to read the blogs: > > > > Anyone win yet? > > > > Peter from Michigan was the winner, with 1.5 VP in the final table. (I > didn't play a game with him myself, so I don't know his last name.) He > played Arika & friends. Congrats to Peter from Michigan. > Jared Strait was #2, with 1.5 VP as well, playing Nos Princes. Damn! 3rd time at the final table....or is it a fourth? I recall Jared qualifying for the finals once but considering ditching it to go play another event. Jared Strait is my hero! > Ben Swainbank was #3 with 1 VP, playing G2-3 Gio with Le Dinh Tho, Gio > allies, & media locations. Wow. A Swainbank / Strait rematch. > 4 & 5 were Josh Duffin (classic law firm) and Stefan Ferenci (cel/CEL guns). > I believe Josh was 4, but I don't know for certain, as I didn't see the > initial seating of the final table. Way to go Josh and Stefan. Congrats to both! Nice "all-star" line-up for the finals. Hope Jeff has this one on DVD. -Robert

[ quoted text not captured ] So, was the two-day format worth it? How many people had qualified for the NAC overall? Did people play the same decks both days? Jeff

Pat wrote: > "Peter D Bakija" <pd...@lightlink.com> wrote in message > news:BF2E730D.215D5%pd...@lightlink.com... > > So hey--for those of us too lazy to read the blogs: > > > > Anyone win yet? > > > > > > > > Peter from Michigan was the winner, with 1.5 VP in the final table. (I > didn't play a game with him myself, so I don't know his last name.) He > played Arika & friends. BAN ARIKA! ;) Gratz to all finalists. Jeff

"Robert Goudie" <rob...@vtesinla.org> wrote in message news:1124674746....@f14g2000cwb.googlegroups.com... > > Pat wrote: >> Jared Strait was #2, with 1.5 VP as well, playing Nos Princes. > > Damn! 3rd time at the final table....or is it a fourth? I recall Jared > qualifying for the finals once but considering ditching it to go play > another event. Jared Strait is my hero! > I asked Jared for a rundown of his final appearances during my day 1 game with him... I believe this was his 6th final table! > Way to go Josh and Stefan. Congrats to both! > > Nice "all-star" line-up for the finals. Hope Jeff has this one on DVD. > It was taped by Oscar & Jeff. (AFAIK, using the same camera; I think they just shared videographer duty.) - Pat

Peter D Bakija wrote: > That is good to hear. Like, full well realizing that having no time limit is > a recipe for madness, but I'd really like, at the very least, the National > Championship not end in a time out. But ya know, it is still a total pipe > dream :-) There's no practical reason why the final can't be raised to 2.5 hours. The game will not just expand like a gas to consume whatever time you throw at it. People just believe that because they'll argue against any change at whatever cost, generally speaking. The status quo is god.

david.che...@gmail.com wrote: > There's no practical reason why the final can't be raised to 2.5 hours. > The game will not just expand like a gas to consume whatever time you > throw at it. People just believe that because they'll argue against > any change at whatever cost, generally speaking. The status quo is god. I'd certainly be in favor of upping, at the very least, the length of finals for, like, national championships or whatever to 3 hours. But then, there are those that would argue that, in fact, games will expand like gas, and that 3 hour finals would time out just as often as 2 hour finals. It would just take longer to time out. I'm not necessarily one of those people--I rarely see games time out in competetive play, but then as I have pointed out elsewhere, I'm what I like to call a "load bearing" player, in that I either win or die trying, which speeds the whole game up for everyone, but when I do see games time out, they rarely are games that are almost over but just run out of time, they are usually games that have hit a stasis wall, and an extra hour would generally result in the game timing out in an extra hour. In any case--congrats to Peter X of Michigan. And special props to ex-Ithaca-home-team-member Joshy boy! [ quoted text not captured ]

Pat wrote: > "Robert Goudie" <rob...@vtesinla.org> wrote in message > news:1124674746....@f14g2000cwb.googlegroups.com... > > > > Pat wrote: > >> Jared Strait was #2, with 1.5 VP as well, playing Nos Princes. > > > > Damn! 3rd time at the final table....or is it a fourth? I recall Jared > > qualifying for the finals once but considering ditching it to go play > > another event. Jared Strait is my hero! > > > > I asked Jared for a rundown of his final appearances during my day 1 game > with him... I believe this was his 6th final table! Yikes. I think my first year there was 1999 so Jared must have grabbed a couple more in the five years prior. Jared is my hero! There are only a few people who've even appeared in 2 finals there. Sure, Jared's been going longer than most but that's still a fine record! -Robert

We've already moved to 2.25 or 2.5 hour final limits here. We'd do the same with prelim rounds if we could get some of the lazy bones players here before noon. Most timed out games would finish with just 15 mins more, or at least break open and hand out some vps. G

[ quoted text not captured ] And, hence, the "dueling assertions" dilemma. Have you noticed a decrease in the number of timeouts for your group of players, since you are actually trying the change out instead of just speculating on it?

<david.che...@gmail.com> wrote in message news:1124716890.3...@z14g2000cwz.googlegroups.com... [ quoted text not captured ] I have nothing against the Continental Finals expanding to 2.5 hours. With the amount of time those guys spent on getting to that point, I don't see the harm of making them work an extra half hour to get a result. But I still disagree that the "the game will not just expand..." to fill the extra half hour. From all descriptions, I get the distinct impression that at least of couple of the past ones would have and other big tournament finals I've seen also have looked like that they got would have gotten nowhere with more time. Often games seem like they're stalled out in some kind of equalibrium situataion until the last 10 minutes, when guys have to start making decisions or the preliminary result tiebreakers will make all the decisions for them. Fred

"Pat" <patrick.l...@comcast.nyetspam.net> wrote > > Peter from Michigan was the winner, with 1.5 VP in the final table. (I > didn't play a game with him myself, so I don't know his last name.) He > played Arika & friends. The winner was Peter Charnley from Ann Arbor, who was also at last year's final table. Cheers, WES

"Andrew 'Wes' Weston" <gh...@NYETSPAMmnsi.net> wrote in message news:dedaj...@enews2.newsguy.com... [ quoted text not captured ] ...and, at this moment, ranked 115th worldwide by our fabulous "ranking" system. Fred

Frederick Scott wrote: > > The winner was Peter Charnley from Ann Arbor, who was also at last year's > > final table. > > ...and, at this moment, ranked 115th worldwide by our fabulous "ranking" > system. ...which will no doubt completely change in a week or so, as that ranking reflects nothing about this NAC yet. The ranking points for the weekend were definitely _not_ entered before we all flew home, as Robyn wasn't given the Archon files from the weekend at the con (and I was sitting beside her when arrangements were being made to email them to her this week.) But without knowing exactly how Peter scored on the qualifier or on Day 1, he's getting at least something on the order of 267 ranking points this weekend (and that assumes he minimally qualified on Day 1) - and that alone is enough to catapult him up the rankings (at least) about 70 places. The bigger (yet still small) concern people had about the rankings is this: both Day 1 and Day 2 count as seperate championship events (because you have to qualify to play in them), and so Andreas gets a few more points for winning Day 1 then Peter does for Day 2 because the field size contracts. > Fred -John Flournoy

On Mon, 22 Aug 2005, Frederick Scott wrote: >> The winner was Peter Charnley from Ann Arbor, who was also at last year's >> final table. > > ...and, at this moment, ranked 115th worldwide by our fabulous "ranking" > system. Well, obviously his rating points for winning the NAC aren't in the system yet. Without those points, he has 505, which is what puts him in 115th place. Assuming about 300 points for winning the NAC this year and a few more for however he did in the first round tournament to advance to the NAC (not sure how he placed, but he ousted me), he should make it right about into the mid-high thirties. That's a fairly respectible ranking. If you look at Peter's overall record, he has ten games not counting this year's NAC and has made final tables in four of them. He's clearly an accomplished player and as soon as his rating points are entered, his name will be listed next to many of the finest in the world. So what's the problem? Matt Morgan

Frederick Scott wrote: > ...and, at this moment, ranked 115th worldwide by our fabulous "ranking" > system. And I bet his rank will jump considerably as a result. Like, if you walk into the World Series of Poker having never played in the event before, and win, you look like a not that good player either. Until you win. [ quoted text not captured ]

[ quoted text not captured ] He's also ranked 27th in the United States by the same fabulous "ranking" system. This stat is more applicable for a discussion of the North American Championship. There are also 8 Canadians ahead of his 505 point rating, so he's roughly 35th in North America. Not too shabby, and certainly not to be dismissed from pulling off a win in a big event. The only weak spot in his tournament record is that he only won a single 8-player event. The big point payoffs are for winning VTES events, not merely making the finals. There is also a large gap between Sept 2004 and March 2005. Presuming he's a student in Ann Arbor (or perhaps a grad who is no longer nearby), it's no wonder that he doesn't have time during the year to do as much VTES as he might like. Whatever. I just wanted to point out that the rankings may not seem as misleading as otherwise portrayed. Jeff

"John Flournoy" <carn...@gmail.com> wrote in message news:1124745505.1...@g47g2000cwa.googlegroups.com... > > Frederick Scott wrote: > >> > The winner was Peter Charnley from Ann Arbor, who was also at last year's >> > final table. >> >> ...and, at this moment, ranked 115th worldwide by our fabulous "ranking" >> system. > > ...which will no doubt completely change in a week or so, as that > ranking reflects nothing about this NAC yet. I wasn't claiming they were. I was just pointing out that 115th is a pretty low ranking for a future Continental Champion - *especially* for one who clearly didn't "come out of nowhere" but was a previous Continental Championship finalist. Worthless. Fred

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF2FCD88.2161B%pd...@lightlink.com... > Frederick Scott wrote: > >> ...and, at this moment, ranked 115th worldwide by our fabulous "ranking" >> system. > > And I bet his rank will jump considerably as a result. > > Like, if you walk into the World Series of Poker having never played in the > event before, bzzzzzztt! You didn't read what Wes said (and confirmed by his profile): he was a finalist the year before. This guy is not a flash in the pan. Fred

<jeff...@pacbell.net> wrote in message news:1124752164.5...@g44g2000cwa.googlegroups.com... > Frederick Scott wrote: >> "Andrew 'Wes' Weston" <gh...@NYETSPAMmnsi.net> wrote in message >> news:dedaj...@enews2.newsguy.com... >> ...and, at this moment, ranked 115th worldwide by our fabulous "ranking" >> system. > > He's also ranked 27th in the United States by the same fabulous > "ranking" system. This stat is more applicable for a discussion of the > North American Championship. I don't know. That could be a reasonable partial explanation if the European players are just better than the Americans. And for various reasons, that might be possible. But another reasonable explanation might also be that many more points are available for packing one's top eight list if there are more and larger and more larger tournaments in Europe concentrated in smaller areas that are easier for the players to commute to. In short, it doesn't matter whether they're better or not - the system will endemically make them *appear* better by its very nature. > There is also a large gap between > Sept 2004 and March 2005. Presuming he's a student in Ann Arbor (or > perhaps a grad who is no longer nearby), it's no wonder that he doesn't > have time during the year to do as much VTES as he might like. Sure. But this is circular to the intent of my post. Obvously, I don't care about some lame excuse for his lack of participation when the issue is about why participation is even being taken into account in the first place. At least, if the ratings are to be taken as a sign of skill. Fred

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Frederick Scott wrote: > "John Flournoy" <carn...@gmail.com> wrote in message > news:1124745505.1...@g47g2000cwa.googlegroups.com... >> >>...which will no doubt completely change in a week or so, as that >>ranking reflects nothing about this NAC yet. > > I wasn't claiming they were. I was just pointing out that 115th is a pretty > low ranking for a future Continental Champion - *especially* for one who > clearly didn't "come out of nowhere" but was a previous Continental Championship > finalist. > > Worthless. Oh, quit your fucking sour-grapes whining, Fred. The ranking system works, whether you like it or not, and whether you're willing to admit it or not. Suck on it, deal with it, and shut the FUCK up about it, OK? YOU GET TO BE WRONG THIS TIME. No rating system can, or ever will, be able to accurately predict future performance -- especially future IMPROVED performance. Stupid-ass. Also, get your facts straight -- he was rated 27th in the US entering the tournament. In fact, here's his full history: Date Type VPs Players Rank Points 04/17/04 CQ 9 48 3 120 06/24/04 Con 1 25 9 9 06/26/04 CQ 0.5 30 DQ 7 07/17/04 Con 3 12 6 25 08/20/04 CQ 0 90 68 5 08/21/04 CC 9 72 4 120 09/18/04 Con 4 19 7 29 03/05/05 Con 9 8 1 105 04/23/05 CQ 3 41 15 25 05/21/05 CQ 8 10 4 72 06/30/05 Con 0 16 10 5 06/30/05 Con 0 24 19 5 Con = Constructed. CQ = Qualifier. CC = Championship. So what do we see? Well, sometimes he does really well, and sometimes he plain old flops -- look at the number of zeros in there! He made most of his rating points from a 3rd place Qualifier finish, a 4th place Championship finish, and a first place standard constructed finish -- but the win was only out of 8th people. The rest of the time, he's pretty much puttered along, not doing very much. Out of 12 tournaments, he's scored less than ten points (1VP or less on the day) in 5 of them, and 30VP or less in three more. What that means is, in eight tournaments he put in a NOTABLY sub-par performance! His average rating point score for a tournament is 43.9 -- NOT at the mondo-high levels that would indicate that he was one of the best players in the USA. Even dropping his worst 4 tournaments (best 8, remember), he still only does good about half the time -- in the bigger ones, where he gets the most points for them. Still think he didn't deserve that 27th-in-the-US rating? You bet your ass he did, because THAT'S HOW HE PERFORMED. Now, this is not meant to bash Peter -- he's obviously a very good player, or he wouldn't have been able to win the Championships this year. And, of course, his rating will reflect this accordingly, along with his performance in any other tournaments over the weekend -- and it will definitely jump up. But frankly, Peter is a PERFECT example of why the rating system DOES, in fact, work. And it works whether you like it or not, Fred. Now get your tinfoil hat out, bitch, and come up with some excuses to explain these COLD, HARD, FACTS away. I'm looking forward to hearing this shit. - -- Derek (By the way, if you compare Ben Peal's (currently #1 in the US) record, you might notice something -- a much more noted lack of sub-10-point performances, and a much higher average performance overall. In fact, Ben's average rating score is 97.5 -- more than twice what Peter's is. And the ratings reflect that, unsurprisingly. Guess they work after all, huh?) -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFDCmi1tQZlu3o7QpERAsRqAKDRUXhnE0IxBdqfrnhPWP4UAcdmcwCfQcqo cc0FM/gJYB1+b7NVHHdOZ9M= =tC2M -----END PGP SIGNATURE-----

Hey guys, as I'm not here often (have to read the group on Google), could anyone send me the decklist of the Giovanni finalist ? As a matter of fact, I'm surprised that a Gio Powerbleed and a Law Firm made it to the finals, seeing as lately I've seen so much weenies and intercepting allies . Or maybe the metagame isn't the same in the US ? Was Stéphane Lavrut there, and how did he do ? Congratulations to all the finalists. Except you Stefan, it's getting sooo tiresome so see you in every Championship finals !! ;) And now I've got the JOL finals to play with Jared and a few other Big Guys, and the French Championship this week-end and... no ready deck !! lol. Maybe I should train for a drinking contest instead ? Cheers from the Grave, Orpheus

"Derek Ray" <lor...@yahoo.com> wrote in message news:hqudnR0pQex...@giganews.com... > The ranking system > works, whether you like it or not, and whether you're willing to admit > it or not. > > Suck on it, deal with it, and shut the FUCK up about it, OK? > YOU GET TO BE WRONG THIS TIME. Actually, that's the only reason I post - because I'm right. > No rating system can, or ever will, be able to accurately predict future > performance -- especially future IMPROVED performance. Huh? I'm confused what the hell is a rating system *FOR* if *NOT* to predict the future? If these ratings prove skill, then someone who wins a Continental Championship ought to be a person who has skill. A ranking system that doesn't predict who will do well isn't worth anything. And the point of posting about this is that clearly he has not "improved" all that much. Last year he was fourth. This year he's first. Big fucking improvement. > Also, get your facts straight -- he was rated 27th in the US entering > the tournament. I *did* have my facts straight: I said he was 115th in the world entering the tournament. I didn't say anything about his US ranking. > So what do we see? Well, sometimes he does really well, and sometimes > he plain old flops -- look at the number of zeros in there! Sure. And by contrast, look at David Tatu's record - who's ranked 4th in the U.S and has over 500 more ratings points (more than double) than Peter. Ooops! There really isn't a lot of contrast there - at least, not in the character of the finishes. David has a number of remarkable finishes - and he has a number of poor ones, too. The main difference is that David has nearly four times as many tournaments as Peter. Of course, I'm cherry picking the example. Others, like Ben Peal clearly do much better on average. But the point is, it doesn't take more consistancy to leave a guy like Charnley in your rear-view mirror with this system. You can do it with just a lot more tournaments and the same kind of performance. > But frankly, Peter is a PERFECT example of why > the rating system DOES, in fact, work. Sorry, this is not "working". > And it works whether you like it or not, Fred. Now get your tinfoil hat > out, bitch, and come up with some excuses to explain these > > COLD, > HARD, > FACTS > > away. With other cold hard facts, what else? Fred

On Mon, 22 Aug 2005, Orpheus wrote: > Was Stéphane Lavrut there, and how did he do ? See for yourself: http://themadnessnetwork.blogspot.com/2005/08/gentlemen-behold.html Matt Morgan

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Frederick Scott wrote: > "Derek Ray" <lor...@yahoo.com> wrote in message > news:hqudnR0pQex...@giganews.com... > >>The ranking system >>works, whether you like it or not, and whether you're willing to admit >>it or not. >> >>Suck on it, deal with it, and shut the FUCK up about it, OK? >>YOU GET TO BE WRONG THIS TIME. > > Actually, that's the only reason I post - because I'm right. Numbers say otherwise, Fred. Suck on it. >>No rating system can, or ever will, be able to accurately predict future >>performance -- especially future IMPROVED performance. > > Huh? I'm confused what the hell is a rating system *FOR* if *NOT* to > predict the future? If these ratings prove skill, then someone who > wins a Continental Championship ought to be a person who has skill. And holy fuck -- he IS! IT WORKS! 27th in the US ain't nothin' to sneeze at, Fred. There's a lot of players in the US. > A ranking system that doesn't predict who will do well isn't worth > anything. And this is why you're wrong. Ratings can only measure performance, and while they provide an indicator of skill, NO RATING SYSTEM EVER can accurately predict the future. People get better while you're not looking, see -- and sometimes, good people play really crap decks and try to win anyway, and do poorly. Sometimes, people get lucky! One name for it is "sandbagging". When gambling is involved, people are doing it to try to hide their true skill. In games like V:TES, people do it just to have fun. Do you really think I brought an !Salubri deck to the last-chance Qualifier at GenCon last year because I thought I was going to WIN with it? Or did I do it so I could make funny tally marks on my badge and hear people say "oh my god, what the fuck are you playing and what do you mean, five agg?" Do you ever wonder why the NCAA college basketball tournament goes to all that trouble to seed games, and then makes them play it out? It's because NO RATING SYSTEM CAN PREDICT THE FUTURE. Look at all the upsets in last year's championships, for example. Nobody would deny that the #1 seeds are better teams than the #12 seeds -- but why didn't the #1 seeds win? Because no rating system can predict the outcome of a single tournament. No rating system can ever account for sandbagging, either. If very skilled players choose to play "questionable" decks, their rating will likely suffer -- or they'll generate a lot of low-point scores. > And the point of posting about this is that clearly he has not > "improved" all that much. Last year he was fourth. This year > he's first. Big fucking improvement. Yeah, actually, that IS a big fucking improvement. >>Also, get your facts straight -- he was rated 27th in the US entering >>the tournament. > > I *did* have my facts straight: I said he was 115th in the world > entering the tournament. I didn't say anything about his US > ranking. You were comparing apples to oranges in a desperate effort to prove something that doesn't exist. Use the facts that MATTER, Fred, not the ones that sound good in your little tinfoil-hat world. World ranking means nothing when discussing performance in a NORTH AMERICAN championship. If they had a "best in North America" list, that would be an excellent indicator. They don't; best we got is the "Xth in the US" list. And oh look! He was 27th. (david tatu complaint paragraph snipped by accident and i'm too lazy to paste it back in; complaint summarized as "Tatu has a shitty average; he only has a good rating because he plays a lot") Yep. If you fire at it often enough, you'll hit the target. But unless you finish well in 8 tournaments, you're not going to EVER move up in rating. It records your best 8 finishes over the last 18 months -- which allows you to play 8 tournaments a year ago, suck in all of them, improve your game, play 8 more, kick ass, and have your TRUE skill shown. And you aren't going to finish well in a tournament without being a good player. You certainly aren't going to do it 8 times without being VERY good, no matter how you try to slice, deny, or distort it. So the guys at the top? They're gonna be good, guaranteed. The guys at the bottom? They gotta prove it. That's the way it is, man. No performance, no points. Again you ignore the most important part; you only have to make and play in a minimum of 8 tournaments to have it count. Tatu's rating IS a reflection of his skill, whether you want to believe it or not -- which even more indicates how accurate the system is. > Of course, I'm cherry picking the example. Because in the face of the facts, it's all you can do. > Others, like Ben Peal clearly > do much better on average. But the point is, it doesn't take more > consistancy to leave a guy like Charnley in your rear-view mirror with > this system. You can do it with just a lot more tournaments and the same > kind of performance. Except it isn't the same kind of performance. Tatu has performed better than Peter over the past 18 months. He has a number of first-place finishes, one of which in a Qualifier; Peter has only one, at an 8-person tournament. He also has a number of final-table finishes ASIDE from the first-place finishes. He has more zeros, but the system is designed to allow you to play "fun" decks and have bad days without totally tanking your rating. Why? Because cream rises to the top. If you never play well, you won't EVER have a good rating. And that's a good thing. Of course, I would expect Peter to catch up to Tatu QUITE a bit after this weekend, as Peter clearly outperformed Tatu. Fortunately, the ratings are designed to accurately reflect performance, and will show this after those records get into the system. Who knew the designers were so clever, huh? Sure, if you don't play, nobody will ever know how good you are. But, Fred, I guarantee you can't show me a single competition in the world where ranking/scoring is NOT directly based on performance, and that's what it comes down to. IF YOU DON'T PLAY; YOU WON'T GET RANK. >>But frankly, Peter is a PERFECT example of why >>the rating system DOES, in fact, work. > > Sorry, this is not "working". Sorry, Freddiekins, it does. The facts show it; the numbers show it; it's all about showin' it here. It's the way it is; it's in your face; you can't beat it; you can't do anything but stick your fingers in your ears and shout LALALALALALALALALA I AM NOT LISTENING TO DEREK NO NO IT CANT BE TRUE NOOOOOO. You know what? Fuck you, Fred. You're a whiny little bitch, and I'm sick of your pissing, moaning, tears, and bullshit. You don't like the rating system because you don't understand it -- and can't get through your head that it DOES work. Face it. It does. Every time we look at individual numbers, we say "yep, yeah, that's about right for him", and then here's good old Freddie, standing out in the rain all red-faced shouting NO NO NO THE NUMBERS DONT MATTER IT DOESNT WORK IT DOESNT WORK. Eventually you just have to close the door and figure he'll come in out of the rain when he gets tired of being all wet. >>And it works whether you like it or not, Fred. Now get your tinfoil hat >>out, bitch, and come up with some excuses to explain these >> >>COLD, >>HARD, >>FACTS >> >>away. > > With other cold hard facts, what else? You mean the total lack of ANY counterexamples WHATSOEVER? That's just what I expect from tinfoil-hat whiny-bitches like you. Noise, tears, and trying to dodge the issue. I have nothing more to say to you; go fuck yourself. - -- Derek insert clever quotation here -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFDCoYltQZlu3o7QpERAqf7AKD1LM8wOQCaaQ8hdyRm/b9BcYyuhACgoFpZ yn4+TEEcIVhefMTSJALxejY= =j/Zz -----END PGP SIGNATURE-----

"Orpheus" <orphe...@free.fr> wrote in message news:1124757522.2...@z14g2000cwz.googlegroups.com... Was Stéphane Lavrut there, and how did he do ? Orpheus He was there but he didnt played. He just played in one draft tournement i think (in which he did quite well if i remember corectly). You should had see him with his "fairy" disguise that he had because he lost a bet... That was soooo funny. Martin

"Derek Ray" <lor...@yahoo.com> wrote in message news:Qs2dndbAIqe...@giganews.com... > Frederick Scott wrote: >> "Derek Ray" <lor...@yahoo.com> wrote in message >> news:hqudnR0pQex...@giganews.com... >> >>>The ranking system >>>works, whether you like it or not, and whether you're willing to admit >>>it or not. >>> >>>Suck on it, deal with it, and shut the FUCK up about it, OK? >>>YOU GET TO BE WRONG THIS TIME. >> >> Actually, that's the only reason I post - because I'm right. > > Numbers say otherwise, Fred. Not hardly. >>>No rating system can, or ever will, be able to accurately predict future >>>performance -- especially future IMPROVED performance. >> >> Huh? I'm confused what the hell is a rating system *FOR* if *NOT* to >> predict the future? If these ratings prove skill, then someone who >> wins a Continental Championship ought to be a person who has skill. > > And holy fuck -- he IS! IT WORKS! > > 27th in the US ain't nothin' to sneeze at, Fred. There's a lot of > players in the US. I guess it depends on what you expect. I'd expect a guy who can make the continental finals to be ranked higher than 27th in his own country. (35th on the continent, since 8 Canadians are also rated higher than him.) Hard to tell how many players are actually active at the moment. >> A ranking system that doesn't predict who will do well isn't worth >> anything. > > And this is why you're wrong. Ratings can only measure performance, and > while they provide an indicator of skill, NO RATING SYSTEM EVER can > accurately predict the future. Not perfectly, of course not. But it should give a good idea of who's more likely to win or it has no purpose. ... > Do you ever wonder why the NCAA college basketball tournament goes to > all that trouble to seed games, and then makes them play it out? It's > because NO RATING SYSTEM CAN PREDICT THE FUTURE. Sure, I understand all that. I'm just saying that a guy who was 4th last year and 1st this year and isn't rated in the top 25 probably isn't being rated properly. I don't know when the last time an NCAA Champion was rated as low as 35th going into the tournament. >> And the point of posting about this is that clearly he has not >> "improved" all that much. Last year he was fourth. This year >> he's first. Big fucking improvement. > > Yeah, actually, that IS a big fucking improvement. No, it's not. It proves that two years in a row, the guy managed to make the finals in a huge tournament where everyone's trying their hardest - no sandbagging. It proves he's not a mediocre player who just happened to get lucky in one tournament. >>>Also, get your facts straight -- he was rated 27th in the US entering >>>the tournament. >> >> I *did* have my facts straight: I said he was 115th in the world >> entering the tournament. I didn't say anything about his US >> ranking. > > You were comparing apples to oranges in a desperate effort to prove > something that doesn't exist. Bullshit. I used the first thing I saw - sending YOUR ass scurrying off to find something that didn't look as bad. I have to admit, I didn't expect to find the placement so lopsided between the U.S. and the rest of the world. But 35th on the continent really doesn't look very good for a guy who can make the finals two years in a row, either. > (david tatu complaint paragraph snipped by accident and i'm too lazy to > paste it back in; complaint summarized as "Tatu has a shitty average; he > only has a good rating because he plays a lot") > > Yep. If you fire at it often enough, you'll hit the target. But unless > you finish well in 8 tournaments, you're not going to EVER move up in > rating. It records your best 8 finishes over the last 18 months -- > which allows you to play 8 tournaments a year ago, suck in all of them, > improve your game, play 8 more, kick ass, and have your TRUE skill > shown. And you aren't going to finish well in a tournament without > being a good player. You certainly aren't going to do it 8 times > without being VERY good, no matter how you try to slice, deny, or > distort it. So the guys at the top? They're gonna be good, guaranteed. > The guys at the bottom? They gotta prove it. That's the way it is, > man. No performance, no points. All of this to divert attention from the fact that your whole big theory of explaining Charnley's rating is bullshit. There's a top 5 player who doesn't do as well on average as he does and that guy has twice Charnley's rating points. > Except it isn't the same kind of performance. Tatu has performed better > than Peter over the past 18 months. He has a number of first-place > finishes, one of which in a Qualifier; Peter has only one, at an > 8-person tournament. Peter has only played in 10 tournaments, not counting the DQ. David has played in 40. And what is the sudden concern about first places? Last post, you were talking up a storm about consistency. Of course the higher ranked guy is going to do better when you look at 1st place finishes - the system emphasizes 1st place finishes. So let's look at comparative consistency, since YOU brought it up. Taking each player's average finishes, by scoring each tournament as function of the number of people who beat them by the total players in that tournament, we find: David: Bloodwork-TotalCon 22/22 0.95 NAQ NorthEast:Boston 40/21 0.50 BloodTears 8/1 0.00 Phobia 9/2 0.11 NAQ GreatLakes:Chicago 48/38 0.77 Friday Night 21/17 0.76 NAQ SouthEast 04 24/9 0.33 Saturday AM PBLA 04 23/10 0.39 NAQ SouthWest 04 24/6 0.21 LA # 4 PM 17/1 0.00 SundayAM LA PB #4 16/11 0.62 Origins Thur Constructed 25/1 0.00 Origins Fri Constructed 25/17 0.64 NAQ Origins 2004 30/24 0.77 WrathOfTheIC 12/1 0.00 NAQ GenCon 2004 90/75 0.82 NorthAmerican Championship 04 72/61 0.82 Dragons Breath 04 24/6 0.21 Fee Stake Atlanta 04 16/5 0.25 Gangrel Revel 11/1 0.00 PrincesHalloweenPizzaTournament 8/4 0.38 ECLastChance04 128/107 0.83 European Championship 04 117/30 0.25 Crocodile's Tongue: TotalCon 24/10 0.38 NAQ North East 05 33/29 0.85 Little Gun, No Fun 9/4 0.33 NAQ SouthCentral 05 20/15 0.70 April Showers 10/5 0.40 Realm Of The BlackSun 16/4 0.19 NAQ South East Atlanta 20/1 0.00 Powerbase: LA 05 19/13 0.63 Powerbase: LA 05 Sat AM 18/14 0.72 Powerbase: LA 05 Sun AM 15/13 0.80 Powerbase: LA 05 Sun PM 14/11 0.71 Origins Friday Const 30/21 0.67 Origins 8pm 14/9 0.57 Origins 4PM 13/6 0.38 Origins 10am 24/14 0.54 NAQ Origins 05 35/26 0.71 Origins Sunday 19/5 0.21 Total 17.74, average fractional finish 0.4435. Peter: NAQ GreatLakes:Chic 48/3 0.04 Origins Thur Constructed 25/9 0.32 Crusade 12/6 0.42 NAQ GenCon 2004 90/68 0.74 NorthAmerican Championship 04 72/4 0.04 Sanguine Instruction 19/7 0.32 Black Lotus 8/1 0.00 NAQ GreatLakes 05 41/15 0.34 Ontario NAQ 2005 10/4 0.30 Origins 2pm 16/10 0.56 Columbus Origins 24/19 0.75 Total 3.83, average fractional finish, 0.3483. In short, David finished higher than about 56% of the other players in the tournaments he played in. Peter finished higher than about 65% of the other players in the tournaments he finished in - not all that impressive for a continental champion, I'll grant. But it hardly explains why he's ranked only 27th in the U.S. when there's a guy who's even far LESS impressive than Peter who's ranked 4th. Your "cold hard facts" do nothing at all to explain away the issue because they don't have anything to do with the issue. The issue is not a matter of consistency. The issue is that scoring people this way doesn't measure skill. > Who knew the designers were so clever, huh? They're not looking so clever, now, huh? Fred

>>> "Orpheus" <orphe...@free.fr> wrote in message news:1124757522.2...@z14g2000cwz.googlegroups.com... Hey guys, as I'm not here often (have to read the group on Google), could anyone send me the decklist of the Giovanni finalist ? As a matter of fact, I'm surprised that a Gio Powerbleed and a Law Firm made it to the finals, seeing as lately I've seen so much weenies and intercepting allies . Or maybe the metagame isn't the same in the US ? <<< I don't think Ben's was a pure traditional power bleed. He had a lot of the same vamps, but with many more allies (Hordes, Ambrosius) and some other tech (WMRH & Channel 10, for example). I believe Ben's NAC deck was very similar to the one of his that's published in the new Player's Guide. Pat

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Frederick Scott wrote: (who cares? same old whiny-ass shit) No, really, Fred. I wasn't kidding. I'm done with you; fuck off. I don't have the time to go over it all twenty times and pound it into your fucked-up bitch skull any more than I already have, in some hope that you might actually listen. So get fucked, OK? Thanks. Bye. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFDCqWmtQZlu3o7QpERAvHxAJ4hqx8xlswft+gEP/7lT7oxWdyaqACfVm22 UnArE9BwKwglnhHaHPMvkU0= =aTYr -----END PGP SIGNATURE-----

Derek Ray wrote: > And it works whether you like it or not, Fred. Now get your tinfoil hat > out, bitch, and come up with some excuses to explain these > > COLD, > HARD, > FACTS > > away. I'm looking forward to hearing this shit. Derek, Will you marry me? [ quoted text not captured ]

Frederick Scott wrote: > bzzzzzztt! You didn't read what Wes said (and confirmed by his profile): he > was a finalist the year before. This guy is not a flash in the pan. Nope. But he was inconsistient in his play--certainly a reasonably player, but not a total ringer. He had won 1 of his previous 12 events on record. Should someone who has won 1 event in the previous 18+ months be any higher that 27th in the US? Unlikely. He has done pretty well, but not won much. 27th in the US/115th in the world sounds perfectly reasonable. He then goes to the national championships with a reasonable, "pretty good" kind of record (as that reflects his status when the event starts) with a good deck, plays well and/or gets lucky. He beats a lot of "better" players and, against the odds, wins. It happens all the time. In all sorts of games. His rating will go up considerably as the result of winning (how many points do you get for winning a 60+ player NAC? 500 or so?) His rating before the NAC reflected a pretty good, if inconsistient player who got into the finals a lot but didn't win much. His rating after the NAC will reflect a pretty good, if inconsistient player who won the NAC through a combination of skill, luck, and that plucky spark that underdogs ride to victory. [ quoted text not captured ]

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Peter D Bakija wrote: > > Derek, > > Will you marry me? I'm sorry, Peter. I'm saving myself for Wes. - -- Derek insert clever quotation here -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFDCx7VtQZlu3o7QpERAtrRAJ9otl8yTNACxI+QtR1SxyQ37kYjsACeJsPH RrA+bYtsEzHTVWM0YWotSU4= =KOzQ -----END PGP SIGNATURE-----

Derek Ray wrote: > I'm sorry, Peter. I'm saving myself for Wes. Man. Wes gets all the breaks. [ quoted text not captured ]

jeff...@pacbell.net wrote: > So, was the two-day format worth it? > How many people had qualified for the NAC overall? > Did people play the same decks both days? Just bumping my questions...with one more... How many people played Assamites? (Props to Tobin so far...:)) Jeff

On Tue, 23 Aug 2005 jeff...@pacbell.net wrote: > jeff...@pacbell.net wrote: >> So, was the two-day format worth it? In what sense? I think it's always worthwhile to have a large tournament to play in. The main reason I heard for the two-day format is that in a 40 player tournament, it's possible to make it to the finals with a single game win, but when that number is doubled, two game wins are required. This forces players to go with high-risk, high-yield decks. Four out of the five finalists had only a single game win so I guess in that sense it was "worthwhile" in that it achieved that goal (assuming that was an actual goal). There was also some talk about keeping the riff-raff out and how this would be the most competative tournament our little continent has ever seen. I saw great players at every table, but the dealing and wheeling and dealing some more was out of hand. >> How many people had qualified for the NAC overall? Dunno, but I think the first round tournament had somewhere between 60 and 70 players. 40 of those advanced to the second tournament. >> Did people play the same decks both days? This year's winner, Peter Charnley, did. Not sure if he tweaked it at all between tournaments. As far as I could tell, most people played different decks both days, but there were a few other exceptions. > Just bumping my questions...with one more... > > How many people played Assamites? (Props to Tobin so far...:)) Just Tobin as far as I know. Matt Morgan

On Tue, 23 Aug 2005, Pat wrote: > I believe Ben's NAC deck was very similar to the one of his that's published > in the new Player's Guide. Actually, it's a very different deck. I'm sure the decklist will be published soon. Ben's NAC finalist deck is mainly a powerbleed that uses Hordes for blocking and combat defense and Le Dinh Tho for annoyance. Matt Morgan

jeff...@pacbell.net wrote: > How many people played Assamites? (Props to Tobin so far...:)) While I have no idea who played what and when at Gen Con, I've seen quite a few Assamites with dominate bleedzookaing while unblockable decks flying around the tournament scene that have been doing quite well overall. [ quoted text not captured ]

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF30A377.2164F%pd...@lightlink.com... > Derek Ray wrote: > >> I'm sorry, Peter. I'm saving myself for Wes. > > Man. Wes gets all the breaks. In life, perhaps, but in VTES that turns out not to be the case. :-) Wes was sadly bubbled out of qualifying in the GenCon last-chance qualifier (he was 6 tournament points behind the last qualifying spot, I think), and if I remember right, also on the bubble just outside of being in the finals of the Shadow Twin draft tournament on Saturday. In this case, consistency is probably not quite what you'd like to have for yourself. ;-) Josh misery loves company - missed the sunday-morning Swainbank-draft finals by 6 TPs myself

<jeff...@pacbell.net> wrote in message news:1124809782.3...@g44g2000cwa.googlegroups.com... > jeff...@pacbell.net wrote: >> So, was the two-day format worth it? > >> How many people had qualified for the NAC overall? Matt answered that one pretty accurately - I heard that there were like 63 people in the Friday "day one" of the NAC, of which 40 made the cut to "day two"; it turned out that you needed 0 GW, 2 VP, and good tiebreakers (or any GW at all) to make that cut, since with only 63 there for day 1, almost two-thirds of the participants were continuing to day 2. >> Did people play the same decks both days? Mostly not, in my experience - let me think - of the people I played with both days (uh, that may have been solely Dave Tatu), none played the same deck (including me); of the ones whose decks I saw both days but didn't play twice with, I can't think of any who played the same deck both days. I did hear of at least two people who did play the same deck (except for some tweaking, possibly) both days, but that's not terribly many out of 40. > Just bumping my questions...with one more... > > How many people played Assamites? (Props to Tobin so far...:)) Yeah, uh, the only one I played against was Tobin, let me think... I can't think of any others off the top of my head from either day of the NAC. They were quite popular in the many (sadly, mostly unsanctioned) drafts though - Black Sunrise and Web of Knives Recruit are totally hot when you're drafting with KMW packs, and the vamps aren't bad either. Josh sorting through piles of email

Joshua Duffin wrote: > <jeff...@pacbell.net> wrote in message > news:1124809782.3...@g44g2000cwa.googlegroups.com... > > jeff...@pacbell.net wrote: > >> So, was the two-day format worth it? Matt also answered this. I guess I mean did those who participated *enjoy* having two play two days worth of VTES for the finals? Was it a "good thing" that it only took 1 GW and many VPs to make the final table after Day 2? Was it successful enough that future continental championships will continue to be modeled after this format? > >> How many people had qualified for the NAC overall? > > Matt answered that one pretty accurately - I heard that there were like > 63 people in the Friday "day one" of the NAC, of which 40 made the cut > to "day two"; it turned out that you needed 0 GW, 2 VP, and good > tiebreakers (or any GW at all) to make that cut, since with only 63 > there for day 1, almost two-thirds of the participants were continuing > to day 2. I actually meant how many people actually qualified, not how many participated at GenCon. I guess it's probably an unknown to players, but perhaps someone from WW knows how many people made the cut prior to the Last Chance Qualifier. > >> Did people play the same decks both days? > > Mostly not, in my experience - let me think - of the people I played > with both days (uh, that may have been solely Dave Tatu), none played > the same deck (including me); of the ones whose decks I saw both days > but didn't play twice with, I can't think of any who played the same > deck both days. I did hear of at least two people who did play the same > deck (except for some tweaking, possibly) both days, but that's not > terribly many out of 40. I guess this is a good thing isn't it? I enjoy playing the same deck 3-4 times in a day, but I doubt I'd want to do it 3-4 times for two days running. I'm gonna really have to plan ahead and do GenCon next year. :) Jeff

<jeff...@pacbell.net> wrote in message news:1124814383.2...@g44g2000cwa.googlegroups.com... > Joshua Duffin wrote: >> <jeff...@pacbell.net> wrote in message >> news:1124809782.3...@g44g2000cwa.googlegroups.com... >> > jeff...@pacbell.net wrote: >> >> So, was the two-day format worth it? > > Matt also answered this. I guess I mean did those who participated > *enjoy* having two play two days worth of VTES for the finals? Was it > a > "good thing" that it only took 1 GW and many VPs to make the final > table after Day 2? Was it successful enough that future continental > championships will continue to be modeled after this format? Oh right, I missed seeing that one on the first line. :-) Hmm... I enjoyed playing two days of championship VTES, but would have also enjoyed (more? hard to say) being able to play a sanctioned draft on the other day (as there ended up being no sanctioned draft on the official GenCon schedule other than the Shadow Twin tournament on Saturday). I do like the "double qualification" idea in concept, in that a 100-player tournament is very different from a 40-player tournament, and you really have to get more lucky to make the finals in the 100-player tournament than the 40 (IMO, keeping in mind that you normally have to get at least somewhat lucky to make the finals even in a 40). But as it actually happened, with 63 players there for day 1? The qualifying requirement for day 2 was (again IMO) relatively minimal, and it seems to me more like you had to get UNlucky to NOT advance in this case, and that a one-day tournament wouldn't have had a hugely different character from the second day in this situation. (Though people probably *would* have chosen different decks, or at least, I might have played my day-2 deck on day 1, though then again, maybe not.) There definitely *should* be less randomness in a two-day format, since in a way you're using six preliminary games to determine the eventual five finalists. But then again, you lose some of the "six game reduced randomness" since the first day is treated as an entirely separate tournament, ie, standings from day 1 are lost, all that matters is whether you made the top-40 cut or not. >> >> How many people had qualified for the NAC overall? >> >> Matt answered that one pretty accurately - I heard that there were >> like >> 63 people in the Friday "day one" of the NAC, of which 40 made the >> cut >> to "day two"; it turned out that you needed 0 GW, 2 VP, and good >> tiebreakers (or any GW at all) to make that cut, since with only 63 >> there for day 1, almost two-thirds of the participants were >> continuing >> to day 2. > > I actually meant how many people actually qualified, not how many > participated at GenCon. I guess it's probably an unknown to players, > but perhaps someone from WW knows how many people made the cut prior > to > the Last Chance Qualifier. Ah, gotcha. By my count there are 78 "qualified before the LCQ" North American-qualified players on White Wolf's list: http://www.white-wolf.com/vtes/index.php?line=Championship. Not quite all of those are people who *live* in North America (eg Stéphane Lavrut), and some NAC players didn't qualify in North America (eg Andreas Nusser), but it should be reasonably close. I seem to vaguely remember that either 9 or 13 people qualified in the LCQ this year, making no more than about 90 possible entrants in NAC Day 1, for about a two-thirds yield of possible vs actual participation. Is that what you meant to wonder? >> >> Did people play the same decks both days? >> >> Mostly not, in my experience - let me think - of the people I played >> with both days (uh, that may have been solely Dave Tatu), none played >> the same deck (including me); of the ones whose decks I saw both days >> but didn't play twice with, I can't think of any who played the same >> deck both days. I did hear of at least two people who did play the >> same >> deck (except for some tweaking, possibly) both days, but that's not >> terribly many out of 40. > > I guess this is a good thing isn't it? I enjoy playing the same deck > 3-4 times in a day, but I doubt I'd want to do it 3-4 times for two > days running. I'm gonna really have to plan ahead and do GenCon next > year. :) Yeah, I did like not having to play the same deck 6 times in a row, although then again, it would be kind of a unique experience, and therefore might be interesting at least once. But we do like to think it's the players we're testing, not just the decks, right? So playing different decks on two days of tournaments makes sense to me in a lot of ways. Josh is suddenly inspired to wonder about a format where no good decks are allowed (something with an extensive banned list, perhaps?) - might turn out to be more boring than when good decks are allowed, though

On Tue, 23 Aug 2005 jeff...@pacbell.net wrote: > Joshua Duffin wrote: >> <jeff...@pacbell.net> wrote in message >> news:1124809782.3...@g44g2000cwa.googlegroups.com... >>> jeff...@pacbell.net wrote: >>>> So, was the two-day format worth it? > > Matt also answered this. I guess I mean did those who participated > *enjoy* having two play two days worth of VTES for the finals? Was it a > "good thing" that it only took 1 GW and many VPs to make the final > table after Day 2? Was it successful enough that future continental > championships will continue to be modeled after this format? Two years ago (my first NAC), I scored 1 GW and 6 VP, which wasn't good enough for that final, but would've put me in tiebreakers for this year. Last year I managed 1 GW and 4 VP, which left me at a distant 20th place. Unfortunately, this year I only got 2.5 VPs with no wins. Guess I'm getting worse. :) Two ways to read that: #1 - I play solid decks that perform above average and can often make the finals, except when the tournament is so huge that I need 2 GW to make the finals. Since I don't go the high-risk, high-yield route, I don't really have much of a shot at a final table when 70-80 players show up. The new 40 player tournament format gives me (and other players like me) a much better chance of making the final. #2 - The competition was much higher this year given the culling process. Had we used this format in previous years, I would not have done as well and still would not have made the final. Probably both of those are true to some degree. At the very least, it gives me some hope of making a final table at some continental championship some time. Anybody want to buy me a ticket to Australia? Matt Morgan

On Tue, 23 Aug 2005, Joshua Duffin wrote: > is suddenly inspired to wonder about a format where no good decks are > allowed (something with an extensive banned list, perhaps?) - might turn > out to be more boring than when good decks are allowed, though Sounds like fun, but holy crap! You'd have to ban Dominate outright and probably also Immortal Grapple, .44 Magnum, KRC, Embrace, 2nd Tradition, WWEF, Auspex, Carrion Crows, Raven Spy, Obfuscate, Jost Werner, Kindred Spirits, Presence, any vampire with capacity below 5...um...er...and a ton of other stuff. Might be easier to let each player use his or her judgment on whether or not his or her deck truly sucks and let the rest of the table "vote him off the island" (i.e. instantly ousted and forfeits any gained VPs) if the deck is too good. Obviously, this would only happen in extreme cases as the table would be giving someone a VP, something that would not be in the interest of most of the players present. The judge could always overrule the vote should ulterior motives be involved (voting one's predator to be ousted because she already has a VP and it will keep one's cross-table buddy in the game). There are a few other stupid variants I'd want to try first. Matt Morgan

Joshua Duffin wrote: > is suddenly inspired to wonder about a format where no good decks are > allowed (something with an extensive banned list, perhaps?) - might turn > out to be more boring than when good decks are allowed, though We could just give everyone one of my decks. ;) Xian

"Derek Ray" <lor...@yahoo.com> wrote in message news:htidnbZi4bg...@giganews.com... > No, really, Fred. I wasn't kidding. I'm done with you; fuck off. I'm sorry you feel that way. But I still feel that the current rating system doesn't reflect player skill very well. Fred

"Xian" <xi...@visi.com> wrote in message news:1124818251.8...@g43g2000cwa.googlegroups.com... [ quoted text not captured ] Ooh! Another interesting format (at least as a thought experiment)! Duplicate VTES: there are 5 different decks at a table; probably each decklist is known in advance to the participants. Each player plays a different deck for each of three rounds (or five rounds if you wanted "total fairness"), probably against different players each round too (if possible). You'd still have shuffling randomness, though, unless the deck were played in the same order each time (and if that order were known, we would probably have gone too far in removing random elements from the game of VTES). Josh ist verruckt!

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF3094DA.21646%pd...@lightlink.com... > Frederick Scott wrote: > >> bzzzzzztt! You didn't read what Wes said (and confirmed by his profile): he >> was a finalist the year before. This guy is not a flash in the pan. > > Nope. But he was inconsistient in his play--certainly a reasonably player, > but not a total ringer. He had won 1 of his previous 12 events on record. > Should someone who has won 1 event in the previous 18+ months be any higher > that 27th in the US? Unlikely. He has done pretty well, but not won much. > 27th in the US/115th in the world sounds perfectly reasonable. > > He then goes to the national championships with a reasonable, "pretty good" > kind of record (as that reflects his status when the event starts) with a > good deck, plays well and/or gets lucky. He beats a lot of "better" players > and, against the odds, wins. It happens all the time. In all sorts of games. You know, I'd agree with you if it was just winning _this_ year's Continental Championship. I'm sure some things are flukes, even two-day 7-round championship events, at least to the extent of allowing someone who isn't quite top of the line caliber to win. But the fact that he was a finalist last year makes it really hard to sell that notion. As I pointed out to Derek, David Tatu is even less consistent in his results yet he's ranked fourth in the U.S. That should tell you that the issue isn't consistency or inconsistency but prolificacy, pure and simple. David plays in lots of tournaments and is ranked much higher than Charnley, who doesn't play nearly as often. That ought to seem silly to people in my book. Fred

Frederick Scott wrote: > But I still feel that the current rating system doesn't reflect > player skill very well. Sure--it is in no way even close to a perfect reflection of player skill. But it is much better that you seem to think it is (given that you think it doesn't at all). [ quoted text not captured ]

In message <BF30A377.2164F%pd...@lightlink.com>, Peter D Bakija <pd...@lightlink.com> writes: >Derek Ray wrote: >> I'm sorry, Peter. I'm saving myself for Wes. > >Man. Wes gets all the breaks. If Derek's the bride, you could be the best man and get all the good bits in the restroom after the ceremony. It's traditional. -- James Coupe PGP Key: 0x5D623D5D YOU ARE IN ERROR. EBD690ECD7A1FB457CA2 NO-ONE IS SCREAMING. 13D7E668C3695D623D5D THANK YOU FOR YOUR COOPERATION.

Frederick Scott wrote: > You know, I'd agree with you if it was just winning _this_ year's Continental > Championship. I'm sure some things are flukes, even two-day 7-round > championship events, at least to the extent of allowing someone who isn't > quite top of the line caliber to win. But the fact that he was a finalist > last year makes it really hard to sell that notion. I'm not saying it was a fluke--he is clearly a "pretty good" player. He got into the finals last year, he won a small local event, and he got some points in qualifiers. That strikes me, in terms of overall performance, as "pretty good". While we don't have hard and fast numbers to work with, I suspect that 27th in the US qualifies as "pretty good". Now that he won the NAC, his score will go up a lot, and he'll move from "pretty good" to "certainly good" (these being ranks I'm making up in my head), as he'll probably get enough points to be top 10 or 15 in the US. Which strikes me as "certainly good" It seems like what we are quibbling about in this particular instance is how good someone is if they get into the finals of the NAC--I think it makes them "pretty good", and Peter's rank bears that out (as 27th in the US strikes me as pretty good). One of the flaws with *any* rating system for VTES is that there are so many variables in any given game, *nothing* is going to be a perfect exemplar of player skill--even in the most complicated ELO system, sometimes you get killed by random cross table shenanigans, or too many Anarch revolts that your grand prey played, or random unlikely vampire contestation, or some iditiot across the table rushing all your guys 'cause he is "role playing" or something. So trying to have an incredibly serious ranking system is counter productive. What we have is a kinda serious ranking system that works pretty well, encourages people to play in tournaments (assuming they care), doesn't punish folks for playing experimental decks, and has a reasonable correlation between high score and being a good player. Does it have scientific accuracy? Not even close. But it is good enough. > As I pointed out to Derek, David Tatu is even less consistent in his results > yet he's ranked fourth in the U.S. That should tell you that the issue isn't > consistency or inconsistency but prolificacy, pure and simple. David plays > in lots of tournaments and is ranked much higher than Charnley, who doesn't > play nearly as often. That ought to seem silly to people in my book. Yet it doesn't. Tatu is certainly a good player--he plays a lot, sure, but to get that high, he needs to play well too. Sure, he might be coasting on 8 good tournaments out of 30, but he is still good enough to do well in those 8 tournaments. If he did poorly in 30 tournaments, he wouldn't be ranked 4th in the US. Which many would look at as a *benefit* of the system--you aren't necessarily punished for playing experimental or risky decks in competition.If someone is super concerned about their rating in an ELO type system, they only ever play the cannon of super good decks--no one ever branches out and experiments with something wacky (which leads to a really stale environment). With the current system (assuming you are super concerned with your rating), if you have a good rating and 8 good games, you can go to an event with something kooky and untested--could be something fantastic, could suck rocks. If you win, you might increase your rating. If you get tooled, your rating isn't hurt. Encourages varried deck play. Which, ya know, I think is good. [ quoted text not captured ]

James Coupe wrote: > If Derek's the bride, you could be the best man and get all the good > bits in the restroom after the ceremony. It's traditional. Score! [ quoted text not captured ]

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF30F568.21687%pd...@lightlink.com... > Frederick Scott wrote: >> You know, I'd agree with you if it was just winning _this_ year's Continental >> Championship. I'm sure some things are flukes, even two-day 7-round >> championship events, at least to the extent of allowing someone who isn't >> quite top of the line caliber to win. But the fact that he was a finalist >> last year makes it really hard to sell that notion. > > I'm not saying it was a fluke--he is clearly a "pretty good" player. He got > into the finals last year, he won a small local event, and he got some > points in qualifiers. That strikes me, in terms of overall performance, as > "pretty good". I guess I'll try and summarize the debate in an attempt to cut off the repetitious circles. What it seems to come down to is, "How far down does a guy have to be on the rating list to raise some eyebrows when he won the Continental Championship this year and was a finalist last year?" The fact that he did this in two consecutive years says much better than "pretty good" to me. This is one of the hardest tournaments in the world to win. One year, OK, maybe a "pretty good" player managed to get over the top. Two years in a row? No way. > What we have is a kinda serious ranking system that > works pretty well, encourages people to play in tournaments (assuming they > care), doesn't punish folks for playing experimental decks, and has a > reasonable correlation between high score and being a good player. Does it > have scientific accuracy? Not even close. But it is good enough. Good enough for what? It's leaving great players way down the list not because they're not great but because they simply don't play enough. The instant you use the rating system to encourage or reward *anything* except good play, it loses its value as a rating system. And, so corrupted, it then loses its value in terms of encouraging the other thing, whatever it is you're trying to use it to encourage. You have to look at what a thing is *for*. A rating system is _for_ giving information, not rewarding behavior you like. If you start bastardizing its information function then it ceases to inform and ultimately does nothing. >> As I pointed out to Derek, David Tatu is even less consistent in his results >> yet he's ranked fourth in the U.S. That should tell you that the issue isn't >> consistency or inconsistency but prolificacy, pure and simple. David plays >> in lots of tournaments and is ranked much higher than Charnley, who doesn't >> play nearly as often. That ought to seem silly to people in my book. > > Yet it doesn't. Tatu is certainly a good player--he plays a lot, sure, but > to get that high, he needs to play well too. Of course. I don't issue with the notion that David's a good player. But how good a player? The reason for bringing up his record was mainly just to counter Derek's suggestion that the reason Peter Charnley is ranked so low has to do with his inconsistency. Whether deliberately or not, David Tatu is a good demonstration of how the system can be "gamed" and that inconsistency truly matters not a bit. In the last thread in which I debated Derek about this, I gave an example of three different ratings formulas that all do the same thing as this system does: reward participation and good results simultaneously. By changing the weightings of different types of rewarded results, I showed how three different players can appear significantly better or worse depending on which formula you chose to use. So how does this tell you anything? If players can slide up and down the ranking list like water depending on the weighting values chosen, what does this say about anything except how well a player scores given the arbitrary formula chosen? Nothing. It doesn't tell you a thing. Fred

"Joshua Duffin" <jtdu...@yahoo.com> wrote in message news:3n12aiF...@individual.net... [ quoted text not captured ] I too think that this format was a success, since the actual championship is going to have overall better games and as a result end up being a better test of skill (I would think and so I was told, I didn't actually make it past the first day.) There are some minor issues I have with the system, although I'm sure those can be worked through (none of the issues playing a factor in why I didn't make it past the first day; these are things I saw that had me wondering.) By the way, someone please remind me not to participate in every draft over the week of nightmares. I generally enjoy draft but five drafts in a week is a tad much for some of us. Albert

Frederick Scott wrote: > I guess I'll try and summarize the debate in an attempt to cut off the > repetitious circles. What it seems to come down to is, "How far down does > a guy have to be on the rating list to raise some eyebrows when he won the > Continental Championship this year and was a finalist last year?" The fact > that he did this in two consecutive years says much better than "pretty > good" to me. This is one of the hardest tournaments in the world to win. > One year, OK, maybe a "pretty good" player managed to get over the top. > Two years in a row? No way. Sure it is a hard tournament. But he also didn't do all that well at most of the other tournaments he went to in the past, what, 18 months--he won a small one, got in some finals here and there, and totally crapped out in just as many as not (again, this is in no way meant so slag on Peter--he is just a fantastic opportunity to discuss the rating system :-)--if he were better than "pretty good", he would have done better *between* the NACs too, and he would havehad a higher rating. So yeah, he got in the finals of the NAC last year. But in between, he didn't do all that well, but tried. This, more that anything to me, illustrates the difficulties of rating VTES at all, rather than illustrating a flaw in the system we have. Compare Peter's performaance to, like, Ben Peal or Matt Morgan--they are rated higher, and consistiently do better. Why did Peter get into the finals of 2 NACs in a row, but not do so hot in the interm? Who knows--maybe he plays goofy decks when not at an NAC. Maybe he just got lucky last time. But if he was better than "pretty good", he would have likely had a higher ranking going into the NAC this year. > Good enough for what? It's leaving great players way down the list not > because they're not great but because they simply don't play enough. So then they should play more. If they can't, well, what are you gonna do? That is why you don't get cash prizes for having a high rating. > The > instant you use the rating system to encourage or reward *anything* except > good play, it loses its value as a rating system. And, so corrupted, it then > loses its value in terms of encouraging the other thing, whatever it is you're > trying to use it to encourage. It encourages people to play (assuming they care about ratings points). Again, the ACBL (American Contract Bridgle League) uses a system that is, for all intents and purposes, virtually identical to the current VTES system. The ACBL is *much* bigger than the VEKN. Everyone is perfectly ok with the idea that someone with a high score either is really good or is ok and plays a lot, and they are ok that while most of the time, a high rating has a reasonable correlation with high skill, sometimes there are oddities. Why is it ok for this very well established, very populated organization but not us? Heck--they even have a daily newspaper collumn. > You have to look at what a thing is *for*. A rating system is _for_ giving > information, not rewarding behavior you like. If you start bastardizing its > information function then it ceases to inform and ultimately does nothing. See, but all concrete examples of "high rating score" = "acceptible aproximation of good play skill" indicates that this isn't the case. The top 10 players in the world, in terms of ranking, are likely the top 10 players in the world, in terms of skill. Yeah, ok, there might be the best player in the world somewhere who never plays in any tournaments, so he has no rating. But that is a flaw with *any* rating system. You need to play to get ranked. > Of course. I don't issue with the notion that David's a good player. But how > good a player? The reason for bringing up his record was mainly just to > counter Derek's suggestion that the reason Peter Charnley is ranked so low > has to do with his inconsistency. Whether deliberately or not, David Tatu is > a good demonstration of how the system can be "gamed" and that inconsistency > truly matters not a bit. Which is fine. If you play in exactly 8 tournaments in 18 months, and do really well in all of them, you get a high rating. There is wiggle room the more tournaments you go to--the more tournaments you play, the more you can screw up. But you still need to do well enough in enough events to keep a high rating. And mediocre players likely can't do that. [ quoted text not captured ]

Albert Chang wrote: > By the way, someone please remind me not to participate in every draft over > the week of nightmares. I generally enjoy draft but five drafts in a week > is a tad much for some of us. Albert! Come back to Ithaca! Why are you still in, where, uh, Texas? [ quoted text not captured ]

In message <xwLOe.70396$DW1.19530@fed1read06>, Frederick Scott <nos...@no.spam.dot.com> writes: >I guess I'll try and summarize the debate in an attempt to cut off the >repetitious circles. What it seems to come down to is, "How far down does >a guy have to be on the rating list to raise some eyebrows when he won the >Continental Championship this year and was a finalist last year?" The fact >that he did this in two consecutive years says much better than "pretty >good" to me. This is one of the hardest tournaments in the world to win. >One year, OK, maybe a "pretty good" player managed to get over the top. >Two years in a row? No way. That's not really the point though. With only 12 tournament performances, several of which show him bombing out, a good showing in the previous nationals, one tournament which he won, what would you rate him as, prior to his win? And how would you derive that from the performance data? He's got one tournament win and a place in a nationals final, so we know he probably isn't a complete dolt. But he's got several wipe-outs, which could be for a number of reasons - erratic player, bad/rushed choice of decks, hosed by a metagame shift he didn't expect ("Hey guys, where did all your bleed decks go? *looks at hand with three bounces in whilst getting killed by KRC*), and so on. V:TES does, however, lend itself somewhat to shock performances. A player can do badly not because of their skill but because of what they choose to play. I've seen a number of players at tournaments I've judged, for example, where I know the player is capable of a lot (I've seen them do it), but they're playing a deck they like. And it's very probably a good deck for what it's trying to do, but that deck style isn't terribly strong, or has an exploitable Achilles heel, or is a deck style they struggle with for some reason. For instance, I've seen one very good player play a very, very straightforward "Rush, Dunk, Repeat" sort of deck - weenie pot/cel, pound, pound, pound. Bombed, because it didn't suit him. I've seen other players tinker with intercept decks for months, but it was often too passive and didn't quite have the oomph to oust when it needed. (Hence my often-made suggestion of supplementary rushes, or a stealth- bleed module, or whatever is necessary to get the oust.) When they've switched decks back to something they're good with, they've done really well. I mean, really, really well. But when a player gets fixated down a personal dead-end (e.g. hacking away at Assamites when you just don't have the knack), any rating system is going to reflect them badly because their outcome is below par. Combine such tendencies with a relatively small number of tournaments to generate a rating from and rating people is hard. [ quoted text not captured ]

John Flournoy wrote: > The bigger (yet still small) concern people had about the rankings is > this: both Day 1 and Day 2 count as seperate championship events > (because you have to qualify to play in them), and so Andreas gets a > few more points for winning Day 1 then Peter does for Day 2 because the > field size contracts. > > >>Fred > > > -John Flournoy > andreas won the last chance qualifier. the day one of the championship was won by myself stefan

On Tue, 23 Aug 2005, James Coupe wrote: > With only 12 tournament performances, several of which show him bombing > out, a good showing in the previous nationals, one tournament which he > won, what would you rate him as, prior to his win? And how would you > derive that from the performance data? We could use an ELO system. Peter would've beaten a number of better-ranked players by getting the NAC final last year and his rating would've soared! Whoa, it really works well, right? Then he'd go back to Ann Arbor and do poorly against all the players who didn't make an NAC final and his rating would completely bomb. He'd have shown up this year with a rock-bottom rating to go on to win and his rating would jump up to the top again. That would be such a great system. Why don't we use that? Matt Morgan

Frederick Scott wrote: > > I wasn't claiming they were. I was just pointing out that 115th is a pretty > low ranking for a future Continental Champion - *especially* for one who > clearly didn't "come out of nowhere" but was a previous Continental Championship > finalist. > > Worthless. > > Fred > > rankings are not supposed to predict the future, they are supposed to evaluate the past (18 month to be precise) peter is an excellent player, but aside from the two finals at gencon he has little to show . (one reason beeing he has not played in that many tourneys, still if he had done in fine in 8 of those tourneys he would be top 10 in the world) stefan

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF3104C0.216A1%pd...@lightlink.com... > Albert Chang wrote: > >> By the way, someone please remind me not to participate in every draft >> over >> the week of nightmares. I generally enjoy draft but five drafts in a >> week >> is a tad much for some of us. > > Albert! Come back to Ithaca! Why are you still in, where, uh, Texas? > Yea, still in Houston trying to graduate and avoid getting fired by my advisor. I never thought I'd say this but boy do I miss undergrad. On another note, with the rotating Championship format and the timing problems it raises, I may try to split my vacation into two half weeks and make both the weekend of Origins and the Championship, when that's set. We'll see how that works out though. [ quoted text not captured ]

Matthew T. Morgan wrote: > Probably both of those are true to some degree. At the very least, it > gives me some hope of making a final table at some continental > championship some time. Anybody want to buy me a ticket to Australia? > > Matt Morgan Come on Matt its just a matter of time till you make the finals at a CC. you are an awesome player and if you have a little luck you´ll be there already in budapest. although you (and everybody else) will need a lot of luck to win in budapest with 200+ players at day 1 and 120+ at day 2. So Oscar, Steve W., Stewart W. , LSJ, Gabor et al, dont you think it would be better to use the (modified probably for a 50 player day 2 event) Lavrut/ Walch system in Budapest, the by far larger attendance at the ec will turn this event into a lottery which will a) create an atmosphere with a lot of bleed decks and degenerate play b) and leave a lot of players dissappointed about the amount of luck needed to get to the finals. i really encourange all of you to think about it. it is never to late to change it. stefan

Frederick Scott wrote: > > I'm not saying it was a fluke--he is clearly a "pretty good" player. He got > > into the finals last year, he won a small local event, and he got some > > points in qualifiers. That strikes me, in terms of overall performance, as > > "pretty good". > > I guess I'll try and summarize the debate in an attempt to cut off the > repetitious circles. What it seems to come down to is, "How far down does > a guy have to be on the rating list to raise some eyebrows when he won the > Continental Championship this year and was a finalist last year?" The fact > that he did this in two consecutive years says much better than "pretty > good" to me. This is one of the hardest tournaments in the world to win. > One year, OK, maybe a "pretty good" player managed to get over the top. > Two years in a row? No way. I'll be repetitive, kinda. How far down does a guy have to be to raise eyebrows? Well, since the ratings are based on _past_ performance, not current, the issue of 'he won this year' is a flawed argument from the get go. "Two years in a row" simply isn't reflected in the rankings yet, so why be surprised when someone's lower ranked? Either you are taking his 2005 win into account, in which case Peter's ranked around at least 40th in the world and roughly 4th in the US. That would not raise the least bit of eyebrows, in my opinion; I'd expect a US player who made two NAC final tables to be in the top 5. Or, you discount it, in which case being in the top 100 or so players worldwide - for having made a Continental finals once and only winning one single tourney in 18 months - isn't out of line either. And keep in mind that the rankings only cover the last 18 months. Apparently, had Jared Strait won, you'd have been even more outraged at his 285th ranking - which reflects neither his being a finalist this year nor having won the NAC in years past. The rankings don't reflect who the 'good players' are, because 'good player' is a nebulous imaginary value. You might as well ask a bunch of magical pixies who the 'good players' are, if you aren't basing it on tangible results. Rankings have to at least partly show who has been doing well (sometimes over a limited period of time) based on tangible results to have any semblance of rationality, and by that definition, Peter was absolutely deserving of his appearing-low-to-you ranking prior to this year's NAC. And his ranking will be appropriately high once this weekend's results are factored in. > Fred -John

Matthew T. Morgan wrote: > On Tue, 23 Aug 2005, James Coupe wrote: > > > With only 12 tournament performances, several of which show him bombing > > out, a good showing in the previous nationals, one tournament which he > > won, what would you rate him as, prior to his win? And how would you > > derive that from the performance data? > > We could use an ELO system. Peter would've beaten a number of > better-ranked players by getting the NAC final last year and his rating > would've soared! Whoa, it really works well, right? Using the old rating system, yes. If the constants had been adjusted properly, one NAC finals appearance would have increased his rating but not made it skyrocket. > Then he'd go back to Ann Arbor and do poorly against all the players who > didn't make an NAC final and his rating would completely bomb. He'd have > shown up this year with a rock-bottom rating to go on to win and his > rating would jump up to the top again. Again, yes, the old system would have done this. However, an ELO system with properly adjusted constants probably would have dropped his rating but probably not enough to make it "rock-bottom". > That would be such a great system. Why don't we use that? While our old ELO system was too volatile it would be unfair to assume that all ELO ratings are necessarily too volatile. Nobody is proposing that we use the old rating system again. -Robert

Stefan Ferenci wrote: > So Oscar, Steve W., Stewart W. , LSJ, Gabor et al, dont you think it > would be better to use the (modified probably for a 50 player day 2 > event) Lavrut/ Walch system in Budapest, the by far larger attendance at > the ec will turn this event into a lottery which will a) create an > atmosphere with a lot of bleed decks and degenerate play b) and leave a > lot of players dissappointed about the amount of luck needed to get to > the finals. > > i really encourange all of you to think about it. it is never to late to > change it. IIRC, there were logistical concerns that made it especially difficult to use this format at the EC this year. I wouldn't be surprised if it were adopted at future ECs, however. -Robert

[ quoted text not captured ] People would get sick by the rapid movement up and down just like in a rollercoaster. -- johannes walch

Frederick Scott wrote: > ... > > Do you ever wonder why the NCAA college basketball tournament goes to > > all that trouble to seed games, and then makes them play it out? It's > > because NO RATING SYSTEM CAN PREDICT THE FUTURE. > > Sure, I understand all that. I'm just saying that a guy who was 4th last > year and 1st this year and isn't rated in the top 25 probably isn't being > rated properly. I don't know when the last time an NCAA Champion was > rated as low as 35th going into the tournament. And of course, going into the championship, Peter wasn't a NAC Champion. Next year, Peter will be entering the tournament ranked very highly, as he should be. > >> And the point of posting about this is that clearly he has not > >> "improved" all that much. Last year he was fourth. This year > >> he's first. Big fucking improvement. > > > > Yeah, actually, that IS a big fucking improvement. > > No, it's not. It proves that two years in a row, the guy managed to > make the finals in a huge tournament where everyone's trying their > hardest - no sandbagging. It proves he's not a mediocre player who > just happened to get lucky in one tournament. Actually, this year's NAC had about 63 players. We had more people at our Qualifier. It's not that huge, even less when you consider that he won the 40-person final - not necessarily the 63-player, seperate-result tourney the day before. And yes, people absolutely sandbagged both tournaments - case in point being one person who brought an expirimental Elihu-Meat Hook deck to the NAC for shits and giggles. > > You were comparing apples to oranges in a desperate effort to prove > > something that doesn't exist. > > Bullshit. I used the first thing I saw - sending YOUR ass scurrying > off to find something that didn't look as bad. I have to admit, I > didn't expect to find the placement so lopsided between the U.S. and > the rest of the world. But 35th on the continent really doesn't look > very good for a guy who can make the finals two years in a row, either. Again, Fred, that's because 35th on the continent reflects a guy who can make it one year in a row. You were too fucking impatient to notice that Peter's relatively low ranking didn't reflect his 2005 NAC results when you looked at it. You goofed, people called you on it, and you insist on continuing to base your arguments around comparing his ranking to performance that isn't part of the rankings yet, and then bitching because the ranking is low. This is why many of us are telling you in varying degrees of politeness that your arguments are spurious. > All of this to divert attention from the fact that your whole big > theory of explaining Charnley's rating is bullshit. There's a top 5 > player who doesn't do as well on average as he does and that guy has > twice Charnley's rating points. > Peter has only played in 10 tournaments, not counting the DQ. David has > played in 40. And what is the sudden concern about first places? Last > post, you were talking up a storm about consistency. Of course the > higher ranked guy is going to do better when you look at 1st place > finishes - the system emphasizes 1st place finishes. So let's look at > comparative consistency, since YOU brought it up. Taking each player's > average finishes, by scoring each tournament as function of the number > of people who beat them by the total players in that tournament, we find: *numbers snipped* > In short, David finished higher than about 56% of the other players in > the tournaments he played in. Peter finished higher than about 65% of > the other players in the tournaments he finished in - not all that > impressive for a continental champion, I'll grant. But it hardly > explains why he's ranked only 27th in the U.S. when there's a guy > who's even far LESS impressive than Peter who's ranked 4th. It's not that impressive for a continental champion, because the NUMBERS YOU ARE QUOTING DO NOT INCLUDE A CONTINENTAL CHAMPIONSHIP, Fred. To use the Tatu-Charnely analogy, when the NAC 2005 is factored in, the two of them will actually be fairly close in rankings, especially in the US where they'll both easily be in the top 10. > Your "cold hard facts" do nothing at all to explain away the > issue because they don't have anything to do with the issue. The > issue is not a matter of consistency. The issue is that scoring > people this way doesn't measure skill. You're right, measuring people this way doesn't measure skill. It can't. And it was never supposed to, despite what you might think. When a player plays a deck strictly for fun that they KNOW is crappy at a tournament, and get a poor result, you'd apparently consider that result to be an indicator of the player's skill. If I play 2 good decks and win 2 tournaments, and deliberately play 10 craptacular wacky decks at another 10 coming in dead last, you apparently would want a ranking system to rate someone who finishes in the top half of those same 12 tournaments higher than me - and then you'd call them better in terms of pure skill. Or consider the other person a 'better player' than me, to use terminology you've previously used. Which is clearly wrong, because my performance in 8 tournaments where I decide to fuck around has zero indicator on my skill or whether or not I'm a 'good player' - and ratings don't measure skill or 'goodness', they measure results, which is the only thing they CAN measure. And good, skillfull players do deliberately fuck around in tournaments all the time, even in large ones. > Fred -John Flournoy

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 John Flournoy wrote: > You're right, measuring people this way doesn't measure skill. It > can't. And it was never supposed to, despite what you might think. Actually, it DOES provide a good ballpark for skill -- because while a skilled player can play poorly and screw up his rating in this system, a weaker player will be completely unable to get a "good" rating. The very best ratings will only be found by winning Qualifiers and Continental Championships, and nobody will deny that it takes a darned good (and possibly slightly lucky) player to get there. And it was even supposed to, to a degree. The design of the system is intentionally such that the players who don't play much, or don't play seriously, will "fall out" of the bottom. While their rating might not reflect their play skill, it won't be because of the system being wrong in some fashion; it'll be because it's their fault for not trying. So, the players at the TOP really _are_ going to be that good -- because you can't get to the top without playing well. And when you come down to it, most people could care less about anything that isn't in the top 20 to 50... so the system is most accurate about measuring what people give a damn about, which is who's the best. > When a player plays a deck strictly for fun that they KNOW is crappy at > a tournament, and get a poor result, you'd apparently consider that > result to be an indicator of the player's skill. If I play 2 good decks The fallacy the current system completely avoids. There is no way to accurately rate players who choose not to play, or who choose to play wacky fun decks -- there just ISN'T. So with the current system, players who choose not to attempt to win end up rated accordingly -- somewhere near the bottom, where it's not necessary to distinguish whether they just plain suck or they're always screwing around. - -- Derek insert clever quotation here -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFDC7cZtQZlu3o7QpERApzSAJ9RQoahtb+GlKG4a+AucDTATeCCCACfeqZd K3S1HhKQLQI02XcFc/meLVI= =mMLk -----END PGP SIGNATURE-----

<jeff...@pacbell.net> wrote in message news:1124809782.3...@g44g2000cwa.googlegroups.com... > jeff...@pacbell.net wrote: >> So, was the two-day format worth it? > >> How many people had qualified for the NAC overall? > >> Did people play the same decks both days? > Jon with the cool shadow box Go Anarch edge played Ahrimanes both days, I think. I don't remember anybody else doing that (besides those mentioned elsewhere in this thread, of course). > Just bumping my questions...with one more... > > How many people played Assamites? (Props to Tobin so far...:)) > > Jeff My predator in the 3rd round of day 1 played Assamites. Tariq & some of the usual suspects from G2. He carved up the table with Devin Villegas's Beast deck. (I believe Devin got the game win, 3-1-1, but it might have been a 2-2-1 tie.) But unfortunately, I don't recall his name. (I really should make notes.) - Pat

"John Flournoy" <carn...@gmail.com> wrote in message news:1124840038....@g49g2000cwa.googlegroups.com... > And of course, going into the championship, Peter wasn't a NAC > Champion. Next year, Peter will be entering the tournament ranked very > highly, as he should be. Given how good he must be to make the finals two years in a row, I am saying that I'd expect him to be rated higher than 35th on the Continent before the tournament. > Frederick Scott wrote: >> >> And the point of posting about this is that clearly he has not >> >> "improved" all that much. Last year he was fourth. This year >> >> he's first. Big fucking improvement. >> > >> > Yeah, actually, that IS a big fucking improvement. >> >> No, it's not. It proves that two years in a row, the guy managed to >> make the finals in a huge tournament where everyone's trying their >> hardest - no sandbagging. It proves he's not a mediocre player who >> just happened to get lucky in one tournament. > > Actually, this year's NAC had about 63 players. We had more people at > our Qualifier. It's not that huge, even less when you consider that he > won the 40-person final - not necessarily the 63-player, > seperate-result tourney the day before. Never mind how large it was. Considering who was playing and what they were playing for, it's clear that it's likely to be much more difficult tournament to win than any mundane 63-player tournament. One would expect at this tournament not to encounter pushovers, experimental decks, nor sandbagging. I suppose it's possible that in spite of the sort of tournament it was, a few people still did such things. But I have a hard time believing it wasn't far rarer than normal. Your evidence to the contrary is anecdotal. > You were too fucking impatient to notice that Peter's relatively low > ranking didn't reflect his 2005 NAC results when you looked at it. You > goofed, people called you on it, Nope. 1) That's not true; of course I knew Robyn hadn't entered the results of the NAC on Monday after it's over. If you believe otherwise, you're stupider than you're accusing me of being. (Hint: this was the significance of the words, "At this moment...") When have the results of any tournament ever shown up on the database 2 days after they happened? Cripes, even Derek wasn't trying to accuse me of not understanding this. and 2) It's got nothing to do with the system needing to have those results entered to credit Peter correctly. I was commenting on the predictive capabilities of a system that works the way this system works. (snip further indignation that I wouldn't cut the system some slack by waiting until it had added in the points from the NAC) > You're right, measuring people this way doesn't measure skill. It > can't. And it was never supposed to, despite what you might think. Then we needn't debate about it and can go on to, "...then what's supposed to whole of point of it, anyway?" Unlike you, Derek holds that it does and you were responding to a post I made in response to one of Derek's posts. > When a player plays a deck strictly for fun that they KNOW is crappy at > a tournament, and get a poor result, you'd apparently consider that > result to be an indicator of the player's skill. If I play 2 good decks > and win 2 tournaments, and deliberately play 10 craptacular wacky decks > at another 10 coming in dead last, you apparently would want a ranking > system to rate someone who finishes in the top half of those same 12 > tournaments higher than me - and then you'd call them better in terms > of pure skill. Or consider the other person a 'better player' than me, > to use terminology you've previously used. I didn't say anything like that. The purpose of the analysis of the Tatu vs. Charnley tournament finishes was to counter Derek's claim that Charnley's low rating was based on his inconsistency. It wasn't - it was due to his lack of prolificacy in attending tournaments. Should tournaments place greater weight on 2 tournament wins or on 10 terrible tournament results stemming from playing wack decks? Neither. Ratings should place equal weight on all games in all tournaments, taking into account the skills of the opponents involved. You might convince me that games in more important tournaments (CQs and continental championships and maybe other types of special tournaments) should have more weight put on their effects on players' ratings - in which case they will increase the gain for winning AND the penalty for losing. But trying to take anything from specific tournament finishes is voodoo science. It's all completely subjective. As for people playing wack decks and doing poorly, so what? If that's the kind of player you are, it should reflect in your ratings. And those ratings should be buffered enough that such finishes are blended in reasonably over time, not taking nosedives when you play weird and careening rapidly when you don't. But ultimately, they're part of your performance and of course they should count. If they don't, it's not an accurate rating. Fred

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF310484.2169F%pd...@lightlink.com... > Sure it is a hard tournament. But he also didn't do all that well at most of > the other tournaments he went to in the past, what, 18 months--he won a > small one, got in some finals here and there, and totally crapped out in > just as many as not (again, this is in no way meant so slag on Peter--he is > just a fantastic opportunity to discuss the rating system :-) (As is David Tatu, by the way. Nice to have guys around with lots of different personalities to use as examples.) > --if he were > better than "pretty good", he would have done better *between* the NACs too, > and he would havehad a higher rating. I think most good players have off-days. Given the low number number of official tournaments, you can't convince me that his lack of other supporting results is anything more than just not having enough chances. I am not trying to argue that he should be up in Peal-Morgan territory. Just that I would expect someone who did that well in two consecutive NACs to be a lot higher than that. The small number of intervening results shouldn't drag him back down so far even if there was much good there - but they did. >> Good enough for what? It's leaving great players way down the list not >> because they're not great but because they simply don't play enough. > > So then they should play more. If they can't, well, what are you gonna do? Um, not penalize them for it. Play and win, that's good. Play and lose, that's bad. Don't play? We can't tell anything from that so we shouldn't. >> The >> instant you use the rating system to encourage or reward *anything* except >> good play, it loses its value as a rating system. And, so corrupted, it then >> loses its value in terms of encouraging the other thing, whatever it is you're >> trying to use it to encourage. > > It encourages people to play (assuming they care about ratings points). > Again, the ACBL (American Contract Bridgle League) uses a system that is, > for all intents and purposes, virtually identical to the current VTES > system. The ACBL is *much* bigger than the VEKN. I don't know anything about the ACBL. There are a lot of potential explanations for such a thing: 1) the system is not *for* rating people's skill, perhaps it's for something else - like qualifying for a higher level of play or something; 2) if the ACBL is much bigger than VEKN, then perhaps it's not a problem to find a tournament - bridge is much more popular than Jyhad, it doesn't require a sizable amount of money and time investment in obtaining cards, and it's quite possible that even people who live in rural Montana can attend lots of tournaments with little effort; or 3) maybe if I knew the situation, I'd disagree with them, too. It seems unlikely to me but I suppose it's possible they have a system that has the same problems as VEKN's system. Maybe the answer is some combination of the three or something else; I'm not in a position to know. >> You have to look at what a thing is *for*. A rating system is _for_ giving >> information, not rewarding behavior you like. If you start bastardizing its >> information function then it ceases to inform and ultimately does nothing. > > See, but all concrete examples of "high rating score" = "acceptible > aproximation of good play skill" indicates that this isn't the case. The top > 10 players in the world, in terms of ranking, are likely the top 10 players > in the world, in terms of skill. I don't agree. It's not clear that this is true - great players may well be missing from the list due to lack of prolificacy. And why be satisfied with only the very best players' ratings? Should an acceptable rating system do a reasonable job of rating all players, given a minimum amount of results for those players? > Yeah, ok, there might be the best player in > the world somewhere who never plays in any tournaments, so he has no rating. > But that is a flaw with *any* rating system. You need to play to get ranked. But you don't necessarily need to play as much as you do to make the current system's top ten list. Fred

"Matthew T. Morgan" <far...@io.com> wrote in message news:2005082316...@eris.io.com... > We could use an ELO system. Peter would've beaten a number of > better-ranked players by getting the NAC final last year and his rating > would've soared! Whoa, it really works well, right? > > Then he'd go back to Ann Arbor and do poorly against all the players who > didn't make an NAC final and his rating would completely bomb. He'd have > shown up this year with a rock-bottom rating to go on to win and his > rating would jump up to the top again. > > That would be such a great system. Why don't we use that? Gosh, that sounds like someone is conparing the old, flawed ELO system to the current one. I guess I have to repeat myself about every three posts for the people who don't follow the debate or don't remember critical points brought up in the past. But so be it. If the coeffecients of the ELO system are set right, no one's ratings should "soar" because of one tournament win, even a Continental Championship. It seems unlikely in the extreme that Peter was an average player before he reached the NAC Finals last year and one would have expected his rating to have reflected that at that the time. His rating should have been very high after the NAC not because he got to the finals of the NAC but because he was GOOD ENOUGH to get to the finals of the NAC. After the NAC, the relative few tournaments in between in which he did poorly shouldn't cause his rating to "completely bomb" and be at rock- bottom on beginning the existing NAC. If they did, I would agree - that would be even worse. The original VEKN ELO system was far worse than the existing system, no question. Fred

Frederick Scott wrote: > (As is David Tatu, by the way. Nice to have guys around with lots of > different > personalities to use as examples.) Agreed. Makes using examples much easier :-) > I think most good players have off-days. Given the low number number of > official tournaments, you can't convince me that his lack of other supporting > results is anything more than just not having enough chances. Wha? He played *twelve* VTES tournaments in 18 months. That strikes me as a perfectly reasonable number of events in 18 months. And probably pretty average for people with mid to high ratings. Yeah, there are some people who play tons of events. But most probably don't. I suspect that the number of people who have high ratings based on large numbers of tournaments (lets call that "The Tatu Factor" :-) is a pretty small percentage of the whole. > I am not > trying to argue that he should be up in Peal-Morgan territory. Just that > I would expect someone who did that well in two consecutive NACs to be a > lot higher than that. And he will be. Previous to winning the NAC, he wasn't someone who did well in two consecutive NACs. He was someone who did well in one consecutive NAC and then crapped out in another 10 tournaments (well, not actually crapped out, but ya know). Once his ratings reflect doing well in two consecutive NACs, he *will* be a lot higher than that (about 4th in the US someone has aproximated). He *had* a rating that reflected doing well in one big event, winning a small event, and then having very uneven performance in a bunch of other events. *Now* he'll have a rating that reflects doing well in two consecutive NACs. > The small number of intervening results shouldn't > drag him back down so far even if there was much good there - but they did. hey didn't drag him down. They just didn't push him up. Doing well in last years NAC alone shouldn't necessarily give him a stellar rating, as the ratings system measures how well you do in 8 tournaments. The NAC helps, as it is weighted heavily, bit as the system measures 8 tournaments, doing well in 1 tournament, even a big one, isn't likely to be that huge. You need to do well in 8 tournaments to have a giant rating. He had a "pretty good" rating. Which makes perfect sense. And will make more perfect sense when he gets rated as top 10 in the US. > Um, not penalize them for it. Play and win, that's good. Play and lose, > that's bad. Don't play? We can't tell anything from that so we > shouldn't. You get "penalized" for not getting to 8 tournaments in 18 months. Which, if you can't get to 8 tournaments in 18 months, is a flaw with the system. But all systems have flaws. > I don't know anything about the ACBL. There are a lot of potential > explanations for such a thing: 1) the system is not *for* rating people's > skill, perhaps it's for something else - like qualifying for a higher level > of play or something; It is as much for rating people's skill as the VEKN system is--kinda sorta not really but with certain correlation. That is what the ACBL system does. It is also what the VTES system does. No one is claiming that the VTES system is a scientific measure of who the best player is. It isn't meant to be. It is, however, meant to measure performance over a number of events. Which it does quite nicely. Luckily, one can draw parallels between having a high rating and being a good player, although there are certainly the occasional outlier. > I don't agree. It's not clear that this is true - great players may well be > missing from the list due to lack of prolificacy. And why be satisfied with > only the very best players' ratings? Should an acceptable rating system do > a reasonable job of rating all players, given a minimum amount of results > for those players? It does. People who play 8 games in 18 months and don't do that well have low ratings. People who play less than 8 games and don't do that well have lower ratings. People who play 8 games and do middle well have middle ratings. People who play 8 games and do very well have high ratings. People who play more than 8 games have more room for error. People who play fewer than 8 games have less room for error and likely lower scores, as the system measures 8 games. It strikes me as doing a good job across the board. Yeah, again, occasionally there is the "Tatu Factor"--someone will play lots and lots of games and not do that well in many of them, but do well enough in enough of them to have a strong rating. But these are likely few and far between. > But you don't necessarily need to play as much as you do to make the current > system's top ten list. You only need to play 8 games in 18 months. You just need to be really good. [ quoted text not captured ]

Frederick Scott wrote: > "Matthew T. Morgan" <far...@io.com> wrote in message > news:2005082316...@eris.io.com... > >>We could use an ELO system. Peter would've beaten a number of >>better-ranked players by getting the NAC final last year and his rating >>would've soared! Whoa, it really works well, right? >> >>Then he'd go back to Ann Arbor and do poorly against all the players who >>didn't make an NAC final and his rating would completely bomb. He'd have >>shown up this year with a rock-bottom rating to go on to win and his >>rating would jump up to the top again. >> >>That would be such a great system. Why don't we use that? > > > Gosh, that sounds like someone is conparing the old, flawed ELO system > to the current one. > > I guess I have to repeat myself about every three posts for the people > who don't follow the debate or don't remember critical points brought > up in the past. But so be it. You maybe shouldn't be surprised by this, considering that probably no more than a dozen people ever studied the old Elo system enough to even understand the points we were trying to make about the coefficients. And that LSJ, a pretty math-savvy guy who probably understood the concepts, didn't agree with us that better coefficients would "fix" the system. > If the coeffecients of the ELO system are set right, no one's ratings > should "soar" because of one tournament win, even a Continental > Championship. It seems unlikely in the extreme that Peter was an > average player before he reached the NAC Finals last year and one would > have expected his rating to have reflected that at that the time. > His rating should have been very high after the NAC not because he got > to the finals of the NAC but because he was GOOD ENOUGH to get to the > finals of the NAC. > > After the NAC, the relative few tournaments in between in which he did > poorly shouldn't cause his rating to "completely bomb" and be at rock- > bottom on beginning the existing NAC. If they did, I would agree - > that would be even worse. The original VEKN ELO system was far worse > than the existing system, no question. Here, though, I'm not sure you're catching the point Matt was making - Peter had few enough rated games (35) in the last 18 months that an Elo system probably wouldn't have enough data to give him a very high rating either. And he won 11 of those 35 games, with 46.5 VPs: higher than the average expectation of 7 games and 35 VPs, but probably not enough higher to yield a very high rating. I'd probably make a separate argument that VTES skill doesn't seem to me to vary all that much anyway - sure, there are bad players, but within the population of good players, the really good ones are mostly not that much better than the moderately good ones. If my impression is accurate (and I don't know that it is, but I do know that I've played with quite a lot of people from all over the place), it's probably not all that interesting to rate people's "true skills" anyhow, since a lot of people will be pretty close to each other - more interesting to watch tournament wins and track recent performance, or something. You know, like the current easily-maintained system does. Yeah, it doesn't actually rate skill like a well-coefficiented Elo system ought to. But that's just not feasible for us, and even if it were, we're not sure that it would be worth the effort. I'd be more interested in postmortems of GenCon games and tournament reports and such than never-ending rating-system discussions, myself, right now. Eh. What can you do... Josh

Matthew T. Morgan wrote: > On Tue, 23 Aug 2005, Joshua Duffin wrote: > >> is suddenly inspired to wonder about a format where no good decks are >> allowed (something with an extensive banned list, perhaps?) - might turn >> out to be more boring than when good decks are allowed, though > > > Sounds like fun, but holy crap! You'd have to ban Dominate outright and > probably also Immortal Grapple, .44 Magnum, KRC, Embrace, 2nd Tradition, > WWEF, Auspex, Carrion Crows, Raven Spy, Obfuscate, Jost Werner, Kindred > Spirits, Presence, any vampire with capacity below 5...um...er...and a > ton of other stuff. Yeah... the banned list would be LENGTHY. Probably almost every card that's the best of its kind or in its discipline. (Quietus might escape unscathed. :-) It would probably be impractical as an actual format. > Might be easier to let each player use his or her judgment on whether or > not his or her deck truly sucks and let the rest of the table "vote him > off the island" (i.e. instantly ousted and forfeits any gained VPs) if > the deck is too good. Obviously, this would only happen in extreme > cases as the table would be giving someone a VP, something that would > not be in the interest of most of the players present. The judge could > always overrule the vote should ulterior motives be involved (voting > one's predator to be ousted because she already has a VP and it will > keep one's cross-table buddy in the game). Yeah, this would be awfully subjective though... almost any deck, no matter how sucky, could look overly strong if it got real lucky, I think. > There are a few other stupid variants I'd want to try first. So many stupid variants, so little time! Josh unduly influenced

Joshua Duffin wrote: > Yeah... the banned list would be LENGTHY. I dunno. You could just make a "playable" list. Hows about something like: The following cards are legal: -Tusk, Talebearer -Appolonius -Tortured Confessions That is all. [ quoted text not captured ]

On Tue, 23 Aug 2005, Peter D Bakija wrote: > Joshua Duffin wrote: > >> Yeah... the banned list would be LENGTHY. > > I dunno. You could just make a "playable" list. Hows about something like: > > The following cards are legal: > -Tusk, Talebearer > -Appolonius > -Tortured Confessions Nice. Fortunately, Appolonius gets a press each combat, so when Tusk (with one blood) blocks his bleed of 2, he can strike hands, press, knock Tusk into torpor and play Tortured Confession. "Hmm...seven Tortured Confession in your hand. Just as I expected!" Matt Morgan

On Tue, 23 Aug 2005, Robert Goudie wrote: > While our old ELO system was too volatile it would be unfair to assume > that all ELO ratings are necessarily too volatile. Nobody is proposing > that we use the old rating system again. Sure, I knew that. I was just having a little fun, but if you cut through the sarcasm, you can pretty well see my point that Josh spells out in another post. Given that Peter only had 10 tournaments on record, he'd either bounce all over the place with his big win, subsequent losses and then even bigger win or he'd be ranked pretty much in the middle still because we wouldn't have enough data yet to adjust his rating appropriately. Matt Morgan

> And yes, people absolutely sandbagged both tournaments - case in point > being one person who brought an expirimental Elihu-Meat Hook deck to the > NAC for shits and giggles. I deny all accusations that I played an Elihu deck built around getting Meat Hooks. . . . Damnit! Um, it wasn't me. It was Feueueuerstein. Yeah, that's it. Ankur Gupta Prince of Lafayette "Where Elihu harvests corn with a scythe at +2 strength."

On Tue, 23 Aug 2005, Joshua Duffin wrote: > I'd be more interested in postmortems of GenCon games and tournament > reports and such than never-ending rating-system discussions, myself, > right now. Eh. What can you do... > > > Josh omfg. Josh posted a response without a tag line. Fred, you ought to be ashamed. He was so depressed he forgot to put in a tag line. Ankur

"Joshua Duffin" <jtdu...@yahoo.com> wrote > > In life, perhaps, but in VTES that turns out not to be the case. :-) Wes > was sadly bubbled out of qualifying in the GenCon last-chance qualifier > (he was 6 tournament points behind the last qualifying spot, I think), and > if I remember right, also on the bubble just outside of being in the > finals of the Shadow Twin draft tournament on Saturday. In this case, > consistency is probably not quite what you'd like to have for yourself. > ;-) But... but... but... I won create-a-clan on Sunday! Honestly, I was less upset about the also-ran thing than everyone else was. People kept coming up to me all weekend and consoling me about it, but I really didn't care all that much. GenCon this year was kind of a last-minute deal for me, so just being able to attend and play some great games was prize enough, as far as I am concerned. Oh, and Derek, I'm sorry, but you're not my type, on account of having a penis. We can still be friends though. Cheers, WES

"John Flournoy" <carn...@gmail.com> wrote > > And keep in mind that the rankings only cover the last 18 months. > Apparently, had Jared Strait won, you'd have been even more outraged at > his 285th ranking - which reflects neither his being a finalist this > year nor having won the NAC in years past. An interesting player to compare to. Jared, as has been pointed out, has made the finals *A LOT*. He is also currently playing in our JOL 2005 tourney final table, which just started yesterday. Clearly, he is an amazing player. But he also plays very rarely. As far as I know, he only really goes to the major events like GenCon and DragonCon and usually kicks some ass when he does. I have seen him at a few minor tournaments, like Storylines and MI vs OH, but those are rare, not to mention irrelevant to the ranking system. So, his ratings in the current system are going to be lower because of low attendance at sanctioned games in the past 18 months. I'm in a similar boat myself, not having the same access to tournaments that I used to. And also because I kind of suck at this game. > You might as well ask a bunch of > magical pixies who the 'good players' are, if you aren't basing it on > tangible results. Can we? I know there was a Changeling LARP at GenCon, dressed in full pixie regalia. They might be free to judge next at year's NAC. > Peter was absolutely deserving of his appearing-low-to-you ranking > prior to this year's NAC. And his ranking will be appropriately high > once this weekend's results are factored in. Peter is somewhat local to me and we have often played together. He is an incredible player and it did not surprise me in the least to see him sitting in the finals again this year. But yeah, based on the criteria used for the current rating system, I'm not surprised that he was that far down either. Cheers, WES

In message <DFPOe.70417$DW1.32905@fed1read06>, Frederick Scott <nos...@no.spam.dot.com> writes: >After the NAC, the relative few tournaments in between in which he did >poorly shouldn't cause his rating to "completely bomb" and be at rock- >bottom on beginning the existing NAC. So, with an ELO system that you like, he'd be... - going quite high but not stellar due to a good NAC showing last year - slipping away somewhat due to a few poor showings - stabilising with a small tournament win and some good showings So, not stellar in the first place. So, maybe it'd take him to somewhere in the top 20 in the US? Then it would be ebbing away slowly. So, maybe 30th in the US? 35th? 40th? Oh look, that'd be roughly where he is now. How would your perfect rating system have ranked him more accurately? Please don't hand-wave. Please show exactly what it would have extracted from available data to get a more accurate ranking. If necessary, invent some plausible other data (such as rough guesses of "Well, let's assume every table has a good player, a better than average player, an average player and a newbie, except at the NAC where everyone is better than average or better...", or similar, if necessary). -- James Coupe PGP Key: 0x5D623D5D YOU ARE IN ERROR. EBD690ECD7A1FB457CA2 NO-ONE IS SCREAMING. 13D7E668C3695D623D5D THANK YOU FOR YOUR COOPERATION.

jeff...@pacbell.net wrote: > I actually meant how many people actually qualified, not how many > participated at GenCon. I guess it's probably an unknown to players, > but perhaps someone from WW knows how many people made the cut prior to > the Last Chance Qualifier. Sure. http://www.white-wolf.com/vtes/index.php?line=Championship (which hasn't been updated from the LCQ yet) -- LSJ (vtesr...@TRAPwhite-wolf.com) V:TES Net.Rep (remove spam trap to reply) Links to V:TES news, rules, cards, utilities, and tournament calendar: http://www.white-wolf.com/vtes/

Matthew T. Morgan wrote: > Nice. Fortunately, Appolonius gets a press each combat, so when Tusk > (with one blood) blocks his bleed of 2, he can strike hands, press, knock > Tusk into torpor and play Tortured Confession. "Hmm...seven Tortured > Confession in your hand. Just as I expected!" Man. I knew there was a loophole in there. This format is broken. :-) [ quoted text not captured ]

"James Coupe" <ja...@zephyr.org.uk> wrote in message news:Ar0jcq0u...@gratiano.zephyr.org.uk... > V:TES does, however, lend itself somewhat to shock performances. A > player can do badly not because of their skill but because of what they > choose to play. I've seen a number of players at tournaments I've > judged, for example, where I know the player is capable of a lot (I've > seen them do it), but they're playing a deck they like. And it's very > probably a good deck for what it's trying to do, but that deck style > isn't terribly strong, or has an exploitable Achilles heel, or is a deck > style they struggle with for some reason. > You know, upon reflection this is actually a (secondary) point in favor of giving Dave Tatu a high rating, despite his larger number of mid-level performances. Dave (and others) achieve their high number of attendances primarily through extensive travel. He's throwing himself into a wide variety of meta-games with little-to-no clue as to what to expect, and so gets "bit" by the meta-game effect often, certainly more often than others who are somewhat more aware of the local meta-trend. The current system allows for this, and doesn't penalize Tatu for his participation in environments that are a total unknown to him. I think thats a positive feature of the current system... DaveZ Atom Weaver

Wes wrote: > But... but... but... I won create-a-clan on Sunday! I would really like to hear an account of the create-a-clan tourney. How did John Flournoy do with the Addams Family? > Oh, and Derek, I'm sorry, but you're not my type, on account of having a > penis. Don't let a little thing like that stop you, Wes.

Emmit Svenson wrote: > Wes wrote: > > But... but... but... I won create-a-clan on Sunday! > > I would really like to hear an account of the create-a-clan tourney. > How did John Flournoy do with the Addams Family? Well enough. 1 VP in round one, ousted by my own Game of Addams/Malkav in the 2nd. Won a prize for 'most fun to play against'. I'll have the artwork posted to the net sometime this week, hopefully, as it got a great response from people who saw it. -John Flournoy

Frederick Scott wrote: > Never mind how large it was. Considering who was playing and what they > were playing for, it's clear that it's likely to be much more difficult > tournament to win than any mundane 63-player tournament. One would expect > at this tournament not to encounter pushovers, experimental decks, nor > sandbagging. I suppose it's possible that in spite of the sort of > tournament it was, a few people still did such things. But I have a > hard time believing it wasn't far rarer than normal. Your evidence to > the contrary is anecdotal. Sure. My only point about that was to note that not everyone was automatically trying their hardest; obviously many of the participants were - and you're right, probably a greater percentage than in other tournies. *some finished debate snipped* > and 2) It's got nothing to do with the system needing to have those > results entered to credit Peter correctly. I was commenting on the > predictive capabilities of a system that works the way this system > works. I'd have to (more politely) agree with Derek here - predictive capabilities are generally poor in any such system. > > You're right, measuring people this way doesn't measure skill. It > > can't. And it was never supposed to, despite what you might think. > > Then we needn't debate about it and can go on to, "...then what's > supposed to whole of point of it, anyway?" Unlike you, Derek holds > that it does and you were responding to a post I made in response > to one of Derek's posts. I'd think that the point is to track performance - certainly performance is in part a result of skill, but there's not a direct 'this ranking system measures skill' correlation. It measures something (as Derek states) that you can't really achieve without a certain amount of skill, but a lack of results doesn't correlate to a lack of skill, so it's not a direct measurement of talent. So at least in that (that the ranking system as written doesn't measure skill) we agree. But I don't see how you can have a ranking system that DOES measure skill over results - given that skill is an incredibly-hard-to-reliably-define attribute absent making assumptions about someone's talent based on their actual results. I'd be happy if there was one, but 'skill' is too subjective. > Should tournaments place greater weight on 2 tournament wins or on > 10 terrible tournament results stemming from playing wack decks? > Neither. Ratings should place equal weight on all games in all > tournaments, taking into account the skills of the opponents > involved. How do you determine the 'skill' of the opponents involved, though? What about the skill of a player in his first tournament, yet who has played Jyhad for 10 years in a competitive environment? Again, you can determine the performance of a player based on his results, but actually seperating skill from other factors (luck, playing his skilled buddy's monster deck, favorable seating, etc) is very, very hard. > You might convince me that games in more important > tournaments (CQs and continental championships and maybe other > types of special tournaments) should have more weight put on their > effects on players' ratings - in which case they will increase the > gain for winning AND the penalty for losing. But trying to take > anything from specific tournament finishes is voodoo science. It's > all completely subjective. Yes, this is exactly my point. Taking anything other than the tangible finishes is indeed voodoo science. And therefore nigh-impossible to track, because it's subjective. No, the ranking system isn't ideal. That's in part because an ideal system (one that allows you to predict future results by comparing relative skills of players) is a nearly impossible one. > As for people playing wack decks and doing poorly, so what? If > that's the kind of player you are, it should reflect in your > ratings. And those ratings should be buffered enough that such > finishes are blended in reasonably over time, not taking nosedives > when you play weird and careening rapidly when you don't. But > ultimately, they're part of your performance and of course they > should count. If they don't, it's not an accurate rating. I agree with this - they certainly should count. My point was, that because they _do_ count, it makes tracking skill and gauging future events still harder to judge. Theoretical example: If Jay Kristoff comes to a local 10-man tournament and plays an all-Aabt Kindred bleed deck for amusement value and zeroes out, his rating will and should factor his poor result into its calculation in some fashion - yet this does not make him a less skilled player, nor any less likely to win another continental championship in the future. So using any ranking system (that would reduce his ranking for such a tourney result) as a predictor of future performance is flawed. (As an aside, the fact that the rankings only track _recent_ performance also make them less useful for future prediction; when a continental championship begins, the ratings already don't reflect the winner from 2 years prior, even though most people would not be surprised by a repeat strong showing regardless of recent results.) > Fred -John Flournoy

John Flournoy wrote: [Addams Family deck] > I'll have the artwork posted to the net sometime this week, hopefully, > as it got a great response from people who saw it. Yeah, that was most excellent. I thought I had done a good job with my crypt and my one library card, but man, those were good. I liked the black & white too, I thought it lent it something extra. :) Xian who didn't even play, due to a math vs. english error

In message <UwPOe.70416$DW1.8756@fed1read06>, Frederick Scott <nos...@no.spam.dot.com> writes: >I don't know anything about the ACBL. There are a lot of potential >explanations for such a thing: 1) the system is not *for* rating people's >skill, perhaps it's for something else - like qualifying for a higher level >of play or something; What do you think most ratings systems do? I play a random game of <something> against you in an ELO system. I lose. You win. My rating score goes down. Your rating score goes up. That's just rated our performances. It hasn't rated our skill, because I could've just screwed up. So we play repeatedly. And our cumulative performance can then be extrapolated as an indicator of skill. But it could just be that (in a game with random chance) I got unlucky a few times. Or possibly that I'm much, much more skilled than you, but you happened to pick a particular vulnerability of mine (a particularly innovative offence, for instance). So, I'm still possibly a better player overall than you, but you beat me. Or maybe you are actually a better player than me, and your score going up is well earned. Almost any rating system which works based on competitive play results is going to rank performance, from which other people extrapolate skill. It ranks "The people who played best over the last year, overall" or whatever. Football (soccer) tables are the same. You win a game, you get 3 points. You draw, you get 1 point. You lose, you get zero points. Doesn't matter whether you're playing the people at the top of the table or the bottom. At the end of the season, it doesn't matter if the people in 2nd and 3rd place beat you in every match, so long as (overall) you won (or drew) enough matches to get more points than them. End result: the team that showed the best performance over the season. When can you rank skill? If, for instance, a sport has an entirely objective scoring system. For example, shooting or archery might be candidates here (if you allow everyone to have the same equipment), or perhaps a track or field event where you measure the time taken or the distance an object is thrown etc. (Take into account wind speed etc., if you really need to be accurate.) But in games where the win/lose results are used, you're going to rank overall performance. The competitive nature means you don't rank objective skill. Giant killings can happen. Grunts can battle through better than finesse, in bad weather. Whatever. [ quoted text not captured ]

In message <qdPOe.70415$DW1.14499@fed1read06>, Frederick Scott <nos...@no.spam.dot.com> writes: >Given how good he must be to make the finals two years in a row, I >am saying that I'd expect him to be rated higher than 35th on the >Continent before the tournament. Why? If the best player IN THE WORLD stopped playing tournaments for a while and (therefore) didn't generate enough useful data, but then came back and won a major tournament, would you point and say "Hah, your rating system doesn't work"? Past performance is not an accurate guide to the future. Shares may go down as well as up. How, precisely, would you extrapolate a much higher rating from the performance data? [ quoted text not captured ]

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF3153C5.216C0%pd...@lightlink.com... > Joshua Duffin wrote: > >> Yeah... the banned list would be LENGTHY. > > I dunno. You could just make a "playable" list. Hows about something > like: > > The following cards are legal: > -Tusk, Talebearer > -Appolonius > -Tortured Confessions > > That is all. I think we could afford to extend the playable list a LITTLE bit. Add: Save Face Blood Bond Cauldron of Blood (nearly a combo with Appolonius!) Advanced Tusk, Talebearer Lazarus Advanced Lazarus Of course, we don't want to make the playable list longer than the banned list would have been, so we might as well stop there. Josh knows how to make a game fun

On Wed, 24 Aug 2005, Joshua Duffin wrote: >> The following cards are legal: >> -Tusk, Talebearer >> -Appolonius >> -Tortured Confessions >> >> That is all. > > I think we could afford to extend the playable list a LITTLE bit. > > Add: > > Save Face > Blood Bond > Cauldron of Blood (nearly a combo with Appolonius!) > Advanced Tusk, Talebearer > Lazarus > Advanced Lazarus No way. Blood Bond is WAY too strong in this format. Since there's no intercept, all Blood Bonds will go unblocked. The first guy everyone will Blood Bond is merged Lazarus. He'll never be able to get into combat to press to the second round and play Cauldron. Replace with Tainted Vitae. That'll give poor Lazarus a better chance to play Tortured Confession on Tusk. Matt Morgan

Matthew T. Morgan wrote: > No way. Blood Bond is WAY too strong in this format. Since there's no > intercept, all Blood Bonds will go unblocked. The first guy everyone will > Blood Bond is merged Lazarus. He'll never be able to get into combat to > press to the second round and play Cauldron. Replace with Tainted Vitae. > That'll give poor Lazarus a better chance to play Tortured Confession on > Tusk. > > Matt Morgan But see, I think the real action goes something like this: Player A: " I get out Tusk!" Player B: "Damn. We contest." Player C: "Ahah! I get out Apolonious!" Player D: "Damn. We contest." Player E: "Eat this! I get out Lazarus!" Player A: "Damn. We contest..." [ quoted text not captured ]

"John Flournoy" <carn...@gmail.com> wrote in message news:1124897275.8...@g47g2000cwa.googlegroups.com... >> and 2) It's got nothing to do with the system needing to have those >> results entered to credit Peter correctly. I was commenting on the >> predictive capabilities of a system that works the way this system >> works. > > I'd have to (more politely) agree with Derek here - predictive > capabilities are generally poor in any such system. It depends on what you expect, I think. I'd expect better than what the existing system did. But since you're not trying to assert the existing system has any (or much), our dispute centers on whether another system (like ELO) might. I guess we'd just be advocating alternative speculations to pursue that argument. >> Then we needn't debate about it and can go on to, "...then what's >> supposed to whole of point of it, anyway?" ... > I'd think that the point is to track performance - certainly > performance is in part a result of skill, but there's not a direct > 'this ranking system measures skill' correlation. It measures something > (as Derek states) that you can't really achieve without a certain > amount of skill, but a lack of results doesn't correlate to a lack of > skill, so it's not a direct measurement of talent. > > So at least in that (that the ranking system as written doesn't measure > skill) we agree. Sure, I suppose it tracks performance - in its particular arbitrary formulaic sort of way. (Different formula chosen, different "performance" tracked.) But what's the point of tracking performance? Some sort of bragging rights? I'm not getting it. What is the point of tracking performance if you can't take anything from it? > But I don't see how you can have a ranking system that DOES measure > skill over results - given that skill is an > incredibly-hard-to-reliably-define attribute absent making assumptions > about someone's talent based on their actual results. I'd be happy if > there was one, but 'skill' is too subjective. The definition of the ELO formula originally used by VEKN attempted to track the likelihood that one player would do better than another in any given game. It seems to me that if you can tell player X is better than player Y intuitively, that ought be trackable through numbers. Otherwise, whenever we hold such opinions, we're probably just fooling ourselves. But I don't think we are. I'm pretty sure Ben Peal is a *lot* better player than I am and I think that would come out consistently in a well-tuned ELO formula. I will admit, I am speculating here - but that's my opinion and I'm sticking to it, until shown otherwise. >> Should tournaments place greater weight on 2 tournament wins or on >> 10 terrible tournament results stemming from playing wack decks? >> Neither. Ratings should place equal weight on all games in all >> tournaments, taking into account the skills of the opponents >> involved. > > How do you determine the 'skill' of the opponents involved, though? > What about the skill of a player in his first tournament, yet who has > played Jyhad for 10 years in a competitive environment? What of it? It's understood that early approximations of skill will be very poor until a certain number of results can be tabulated. It's also clear that there are flaws in the theory that no system can correct for - such as a player playing poorly for a year, withdrawing for a period of time and playing only non-rated play and becoming much better, and then rejoining rated play. Nothing you can do about such things but accept that no system is perfect. But you have to try to track skill from past performance reasonably to have a shot at creating a reasonably good system. > Again, you can determine the performance of a player based on his > results, but actually separating skill from other factors (luck, > playing his skilled buddy's monster deck, favorable seating, etc) is > very, very hard. Well, you don't do it at all. Most of those factors are luck and the others are player proclivities - that is, tendencies to do things that are more or less consistent over time. (If I tend to play experimental whack decks 20% of time, presumably that tends to be consistent.) Luck is boiled out by tuning the coefficients down to levels where a string of lucky or unlucky games will generally average out, at the cost of requiring more results to converge the player's rating onto his true skill level. >> You might convince me that games in more important >> tournaments (CQs and continental championships and maybe other >> types of special tournaments) should have more weight put on their >> effects on players' ratings - in which case they will increase the >> gain for winning AND the penalty for losing. But trying to take >> anything from specific tournament finishes is voodoo science. It's >> all completely subjective. > > Yes, this is exactly my point. Taking anything other than the tangible > finishes is indeed voodoo science. And therefore nigh-impossible to > track, because it's subjective. I'm not sure what point you're trying to make here. My conclusion from this point is that tournament finishes should be ignored - only game results should be used. >> As for people playing whack decks and doing poorly, so what? If >> that's the kind of player you are, it should reflect in your >> ratings. And those ratings should be buffered enough that such >> finishes are blended in reasonably over time, not taking nosedives >> when you play weird and careening rapidly when you don't. But >> ultimately, they're part of your performance and of course they >> should count. If they don't, it's not an accurate rating. > > I agree with this - they certainly should count. My point was, that > because they _do_ count, it makes tracking skill and gauging future > events still harder to judge. > > Theoretical example: If Jay Kristoff comes to a local 10-man tournament > and plays an all-Aabt Kindred bleed deck for amusement value and zeroes > out, his rating will and should factor his poor result into its > calculation in some fashion - yet this does not make him a less skilled > player, nor any less likely to win another continental championship in > the future. Yea, but it makes him less likely to win small tournaments. And the skill rating doesn't say anything about which tournaments a player is more or less likely take seriously. So if you only counted his results at Continental Championships, that would cause his rating to be overestimated at small tournaments. I do sympathize with what you're saying, though. If you tried to rate pro football teams based on their pre-season games as well as their regular season games, you obviously wouldn't get as good and estimation of their abilities. I don't think this is quite as bad a situation though (maybe more like trying to use a football team's game results after they've already sewn up the division title and are only playing for playoff seeding) and I see it as worthwhile tracking anyway. However, thinking about such things is bringing me to the opinion that things like qualifiers should have significantly greater weight, even in an ELO system, than mundane tournaments. Fred

"James Coupe" <ja...@zephyr.org.uk> wrote in message news:v2C47$QwSMD...@gratiano.zephyr.org.uk... > In message <qdPOe.70415$DW1.14499@fed1read06>, Frederick Scott > <nos...@no.spam.dot.com> writes: >>Given how good he must be to make the finals two years in a row, I >>am saying that I'd expect him to be rated higher than 35th on the >>Continent before the tournament. > > Why? > > If the best player IN THE WORLD stopped playing tournaments for a while > and (therefore) didn't generate enough useful data, but then came back > and won a major tournament, would you point and say "Hah, your rating > system doesn't work"? You're assuming that if someone stops playing, you don't use past data. Why are you assuming that? The current system has to do something like that because it insists on using participation as part of its rating and must therefore slide the time period from which it considers results along with the date. (Past 18 months, in this case.) If you don't construct your system that way, you're under no obligation to throw out old results just because they're old. > Past performance is not an accurate guide to the future. I'm assuming it is. If not, I eagerly await the day I happen to be - by pure chance - the best player in the world. Like monkeys trying to type out Shakespeare's works, at some point it should eventually happen. Fred

On Wed, 24 Aug 2005 17:34:14 -0400, Peter D Bakija <pd...@lightlink.com> scrawled: >Matthew T. Morgan wrote: > >> No way. Blood Bond is WAY too strong in this format. Since there's no >> intercept, all Blood Bonds will go unblocked. The first guy everyone will >> Blood Bond is merged Lazarus. He'll never be able to get into combat to >> press to the second round and play Cauldron. Replace with Tainted Vitae. >> That'll give poor Lazarus a better chance to play Tortured Confession on >> Tusk. >> >> Matt Morgan > >But see, I think the real action goes something like this: > >Player A: " I get out Tusk!" >Player B: "Damn. We contest." >Player C: "Ahah! I get out Apolonious!" >Player D: "Damn. We contest." >Player E: "Eat this! I get out Lazarus!" >Player A: "Damn. We contest..." I think that last one would be more like: Player A: "Phew! I have advanced La...Damn. We contest..." salem http://www.users.tpg.com.au/adsltqna/VtES/index.htm (replace "hotmail" with "yahoo" to email)

On 24 Aug 2005 08:47:38 -0700, "Xian" <xi...@visi.com> scrawled: > >John Flournoy wrote: >[Addams Family deck] >> I'll have the artwork posted to the net sometime this week, hopefully, >> as it got a great response from people who saw it. > >Yeah, that was most excellent. > >I thought I had done a good job with my crypt and my one library card, >but man, those were good. I liked the black & white too, I thought it >lent it something extra. :) post your's too, then. :) [ quoted text not captured ]

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF314730.216BA%pd...@lightlink.com... > Frederick Scott wrote: >> I think most good players have off-days. Given the low number number of >> official tournaments, you can't convince me that his lack of other supporting >> results is anything more than just not having enough chances. > > Wha? He played *twelve* VTES tournaments in 18 months. That strikes me as a > perfectly reasonable number of events in 18 months. And probably pretty > average for people with mid to high ratings. Apparently, you didn't see or else didn't understand the import of the Tatu- vs.-Charnley analysis. David's average tournament finish turns out to be much lower than Peter's, so David's much higher rating is based purely on his prolificacy. (Derek, at one point, also suggested Charnley's lack of 1st place finishes might also figure in but I think you could take any single non-first-place Charnley and turn it into a first place finish - a "conjecturally improved Charnley" - except possibly the 2004 Continental Championship, which Tatu never won either, and it wouldn't improve Charnley's rating and rankings that much. One additional first place finish would make Charnley's proportion of them greater than Tatu's. This demonstrates to me that first place finishes aren't the issue, either.) David has off-days, Charnley has off-days. David is rated fourth in the country, Charnley is rated 27th. Off days do not kill a rating. Lack of tournament attendance is the only reason Peter is rated so low. >> I am not >> trying to argue that he should be up in Peal-Morgan territory. Just that >> I would expect someone who did that well in two consecutive NACs to be a >> lot higher than that. > > And he will be. Previous to winning the NAC, he wasn't someone who did well > in two consecutive NACs. No, but previous to winning the NAC, he was pretty much just as good as he is now. The existing system failed to show that because of its endemic bias against players who haven't played as much recently. Other systems may well be able to show it as long as Peter Charnley had a sufficient number of results since beginning tournament play - whenever that was. >> The small number of intervening results shouldn't >> drag him back down so far even if there was much good there - but they did. > > They didn't drag him down. The comparison between Peter and David demonstrates otherwise. >> Um, not penalize them for it. Play and win, that's good. Play and lose, >> that's bad. Don't play? We can't tell anything from that so we >> shouldn't. > > You get "penalized" for not getting to 8 tournaments in 18 months. You get penalized for not going to as many tournaments as possible and thus giving yourself more chances to increase your top eight. And, clearly, playing in lots of tournaments has that effect. Granted, as the number goes higher and higher, the chances of helping yourself decreases gradually but playing another tournament is always a good thing. Thus, not playing is a penalty by comparison. Frame it any way you want, it boils down to that. > Yeah, again, occasionally there is the "Tatu Factor"--someone will play lots > and lots of games and not do that well in many of them, but do well enough > in enough of them to have a strong rating. But these are likely few and far > between. I don't know why you suggest that. With all due respect, Peter, I think you really aren't understanding how important the Tatu Factor is. There may be times it's much less important, as with the very best players once they've attended and done very well in eight sizeable tournaments. But usually, I'm pretty sure it makes a huge difference. Fred

"James Coupe" <ja...@zephyr.org.uk> wrote in message news:8Wz8jHQ2...@gratiano.zephyr.org.uk... > In message <UwPOe.70416$DW1.8756@fed1read06>, Frederick Scott > <nos...@no.spam.dot.com> writes: > Almost any rating system which works based on competitive play results > is going to rank performance, from which other people extrapolate skill. > It ranks "The people who played best over the last year, overall" or > whatever. > > Football (soccer) tables are the same. You win a game, you get 3 > points. You draw, you get 1 point. That's not a rating system. Such tables are not *for* rating and no one assumes they are, except in a very crude way. Such things are used to determine the results of "championship season" by the league, or to determine the playoff format leading to the resolution of the chamionship season. This is total apples and oranges. (I'm not going to address the "performance vs. skill" ranking comments as I'd just make the exact same points in response to you as in response to John Flournoy, to Derek, and more directly in this sense to Josh Duffin.) Fred

"Matthew T. Morgan" <far...@io.com> wrote in message news:2005082322...@eris.io.com... > On Tue, 23 Aug 2005, Robert Goudie wrote: > >> While our old ELO system was too volatile it would be unfair to assume >> that all ELO ratings are necessarily too volatile. Nobody is proposing >> that we use the old rating system again. > > Sure, I knew that. I was just having a little fun, but if you cut through the sarcasm, you can pretty well see my point that Josh > spells out in another post. Given that Peter only had 10 tournaments on record, he'd either bounce all over the place with his > big win, subsequent losses and then even bigger win or he'd be ranked pretty much in the middle still because we wouldn't have > enough data yet to adjust his rating appropriately. Actually, if he only did have 10 tournaments (actually it's 11 - I counted wrong; 12 with the DQ) since the inception of his tournament play, the the second thing would be correct. He'd still be slogging up from the middle of the pack. But no one said he had only 12 tournaments in the whole of his careeer - only 12 in the last 18 months. To conjecture that these are the only 12, you'd have to assume a third place finish in his first CQ was his first tournament ever. Does that seem very likely to you? I suppose it's possible. Maybe the guy's just a CCG-playing genius and/or that he'd played enough pickup games that he was ready to burst forth and finish 3rd in a 48-player CQ followed by a 4th in a Continental Championship - then just fiddle-farted around for a year until returning and winning the next Continental Championship. It's possible. But my faith in ELO as a system has to do with assuming that such things are highly unlikely and that prior to doing so well last year, Peter Charnley probably demonstrated his skill in prior tournaments. In ELO, you can use these results because results don't go bad after a period of time. But the current system insists on rating people's attendence as well as their skill - and hence must throw otherwise useful information out. Fred

Frederick Scott wrote: > Apparently, you didn't see or else didn't understand the import of the Tatu- > vs.-Charnley analysis. David's average tournament finish turns out to be > much lower than Peter's, so David's much higher rating is based purely > on his prolificacy. No, no--I got that. I'm not saying that The Tatu Factor can't be a factor. I'm saying that the number of actual participants that it pushes into an unusually high ranking is likely low enough for them to be considered acceptible outliers to the system. > David has off-days, Charnley has off-days. David is rated fourth in the > country, Charnley is rated 27th. Off days do not kill a rating. Lack > of tournament attendance is the only reason Peter is rated so low. Well, was rated so low. He is now going to be rated high. But in any case, yeah, I know that attendance can be a factor in a high vs lower rating. But my point was simply that there is like 100 Charnleys for every 1 Tatu. I'm ok with that. > No, but previous to winning the NAC, he was pretty much just as good as he > is now. He was, but his performance wasn't. The rating system doesn't measure inherrent skill. It measures tournament performance. It doesn't claim to measure inherent player skill. It only claims to measure performance. Previous to winning the NAC, he had pretty good performance. Now he has certainly good performance. And a higher rating to match. > I don't know why you suggest that. With all due respect, Peter, I think you > really aren't understanding how important the Tatu Factor is. No, no--I am. I'm just pretty well convicned that in terms of the overall ratings across the board, the Tatu Factor is pretty insignificant--again, there is likely, say, 1 Tatu (someone whose rating is very high compared to their overall average game performance based on a large number of tournaments) for every 100 Charnleys (someone whose rating is pretty much in line with their overall game performance based on a smaller number of games). Yes. The Tatu Factor can be very significant in terms of an individual rating. But I suspect, again, that it really only effects a very small number of players in the system. A few people play tons of events. Most people play an average number of events. Peter D Bakija pd...@lightlink.com http://www.lightlink.com/pdb6 "So in conclusion, our business plan is to sell hot, easily spilled liquids to naked people." -Brittni Meil

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF32B6A7.21729%pd...@lightlink.com... > Frederick Scott wrote: > >> Apparently, you didn't see or else didn't understand the import of the Tatu- >> vs.-Charnley analysis. David's average tournament finish turns out to be >> much lower than Peter's, so David's much higher rating is based purely >> on his prolificacy. > > No, no--I got that. I'm not saying that The Tatu Factor can't be a factor. > I'm saying that the number of actual participants that it pushes into an > unusually high ranking is likely low enough for them to be considered > acceptible outliers to the system. IMHO, it's much more important. That everyone everywhere on the board is affected by it, and could be rated and ranked much higher or much lower if they were to attend many more tournaments or many fewer tournaments. The main exceptions, IMO (and these would still be affected, depending on which tournaments you're talking about) would be some of the very highly rated players who wouldn't move much if they added or dropped certain tournament. I don't see why your estimation of it is limited to considering guys at the very top. >> I don't know why you suggest that. With all due respect, Peter, I think you >> really aren't understanding how important the Tatu Factor is. > > No, no--I am. I'm just pretty well convinced that in terms of the overall > ratings across the board, the Tatu Factor is pretty insignificant--again, > there is likely, say, 1 Tatu (someone whose rating is very high compared to > their overall average game performance based on a large number of > tournaments) for every 100 Charnleys (someone whose rating is pretty much in > line with their overall game performance based on a smaller number of > games). Yes. The Tatu Factor can be very significant in terms of an > individual rating. But I suspect, again, that it really only effects a very > small number of players in the system. A few people play tons of events. > Most people play an average number of events. Well, I guess we're going to have to disagree about that. In my book, it's the Ben Peals who can place first or very highly in most of the tournaments they enter who are the rare exceptions and the kind of people you can say aren't affected so much by their level of participation. Charnleys and Tatus are probably much more common - especially as you get down off the very top of the list. And I also suspect we disagree about how sensitive the rating is to the exact level of participation. It looks like it's pretty sensitive to me, even once you get past 8 tournaments. Fred

"Emmit Svenson" <emmits...@hotmail.com> wrote > > I would really like to hear an account of the create-a-clan tourney. > How did John Flournoy do with the Addams Family? His deck was simply incredible. The art from the TV show (which I have personally never seen) was all very thematic. For example, Telepathic Misdirection had an image of a disembodied hand pointing to the left. John, if you're listening in, I'd love to see the art posted somewhere. As for a report, I can do that sure, but I'll do it in a separate thread. I did send my clan decklist to Eric Simon, thinking he might want to use it in his anarch newsletters, but he has not responded, so I'll just go ahead and post it myself. Cheers, WES

In message <Et8Pe.70648$DW1.47979@fed1read06>, Frederick Scott <nos...@no.spam.dot.com> writes: >But no one said he had only 12 tournaments in the whole of his careeer - >only 12 in the last 18 months. To conjecture that these are the only >12, you'd have to assume a third place finish in his first CQ was his >first tournament ever. Does that seem very likely to you? If you're going to argue that we should maintain everyone's ratings forever, you have two issues: 1) Impracticality. Robert Goudie has pointed out just how difficult it is to continually recalculate ratings when errors are found. (Duplicate membership numbers, or whatever.) 2) That the game changes. It is entirely possible for someone to have been plugging away at a given strategy, but only to finally push it over the edge on the release of a given set. Say, you want to play Setite Corruption strategies. Results prior to the Final Nights (or KMW, or...) set might have you underperforming, because you're plugging away at a strategy that doesn't work so well. Then, oh look, the perfect card turns up for you and bang, you're away. How good someone was at playing in the environment three years ago is not the same as how good they are now. Bear in mind, for instance, that many political players used table-seating changing votes as a matter of course and got many, many VPs they wouldn't get now as a result. So why would we want their performance with a vastly different card set to influence their rating now? Certainly, a time limit is needed, but expiring their performance over time reflects the realities of V:TES. And, I've also met several good players who simply don't like tournaments - they don't like the places they're held, they don't like giving up all day to it, or whatever. But some of them are very, very good players. Lady Legbiter would be one example - very good player, doesn't like tournaments. What would you rating system do when players like that choose to play? -- James Coupe PGP Key: 0x5D623D5D YOU ARE IN ERROR. EBD690ECD7A1FB457CA2 NO-ONE IS SCREAMING. 13D7E668C3695D623D5D THANK YOU FOR YOUR COOPERATION.

Frederick Scott wrote: > IMHO, it's much more important. That everyone everywhere on the board is > affected by it, and could be rated and ranked much higher or much lower if > they were to attend many more tournaments or many fewer tournaments. The > main exceptions, IMO (and these would still be affected, depending on which > tournaments you're talking about) would be some of the very highly rated > players who wouldn't move much if they added or dropped certain tournament. > I don't see why your estimation of it is limited to considering guys at > the very top. 'Cause it is likely only a few people (mabe I'll get off my ass one day and go look at the records of, like, the top 20 or 50 or something and see where that stands--what were the useful indexes? The ratio of VP's gained vs Games Played?) in the whole system have ratings that are really high when their overall performance is middling due to a huge nuber of events attended. Which means that the numbers are only skewed (if that is how you want to look at it) by a couple points--the guy who is "player number 10" in the US might realistically be player number 11 or 9, but for a bit of, essentially, rounding error. As the system is not meant to be scientifically accurate, that strikes me as acceptible. Like, yeah, if all of the top 10 players in the world, or whatever, were there on the Tatu Factor (and again, this term is in no way meant as a slag on David Tatu in any way, shape, or form--he is a great player who plays a lot of tournaments with uneven results as he likes playing experimental decks in competition, It seems likely that if he played proven archetypes in every event, he'd be the number one player in the world. By a lot :-). But they aren't. The top 10 players in the world are all there by "conventional" means (they play a lot and do well a lot). It seems likely that, again, the *vast* majority of the top 50, if not the whole system, is the result of conventional ratings, rather than Tatu ratings. Which I think is ok, even if a few ratings are compromised, or whatever. > Well, I guess we're going to have to disagree about that. In my book, it's > the Ben Peals who can place first or very highly in most of the tournaments > they enter who are the rare exceptions and the kind of people you can say > aren't affected so much by their level of participation. Charnleys and > Tatus are probably much more common - especially as you get down off the > very top of the list. And I also suspect we disagree about how sensitive > the rating is to the exact level of participation. It looks like it's > pretty sensitive to me, even once you get past 8 tournaments. Maybe. I'll go look at the numbers. I think it is safe to assume that a ratio of VPs/Games is a reasonable indicator of some kind of "skill" measurement. I'll crunch some numbers and see what comes up. [ quoted text not captured ]

"Frederick Scott" <nos...@no.spam.dot.com> wrote in message news:kE7Pe.70483$DW1.3840@fed1read06... > > "James Coupe" <ja...@zephyr.org.uk> wrote in message > news:v2C47$QwSMD...@gratiano.zephyr.org.uk... >> In message <qdPOe.70415$DW1.14499@fed1read06>, Frederick Scott >> <nos...@no.spam.dot.com> writes: >>>Given how good he must be to make the finals two years in a row, I >>>am saying that I'd expect him to be rated higher than 35th on the >>>Continent before the tournament. >> >> Why? >> >> If the best player IN THE WORLD stopped playing tournaments for a while >> and (therefore) didn't generate enough useful data, but then came back >> and won a major tournament, would you point and say "Hah, your rating >> system doesn't work"? > > You're assuming that if someone stops playing, you don't use past > data. Why are you assuming that? > > The current system has to do something like that because it insists > on using participation as part of its rating and must therefore slide > the time period from which it considers results along with the date. > (Past 18 months, in this case.) If you don't construct your system > that way, you're under no obligation to throw out old results just > because they're old. > You've raised this point before(ad nauseam)... you seem to be neglecting an aspect of VTES that to me seems both obvious and critical. Unlike chess, bridge, or any of the other rated games used as a comparitor in this thread, VTES (both its rules and its play components) changes over time. It could just as easily be said that the rating system slides the time period (very slowly) because the game of VTES today is significantly different from the game of VTES played 18 months ago. Perhaps a reason you, in fact, _must_ use participation and a sliding time frame in rating VTES players is bacause, if you want to rate player skill in a VTES game today or next week, then the results from some longer time period ago aren't that valuable a predictor of future performance. Go back any given 18 month timeframe in WW-era VTES, an you end up excluding between one and three expansions from play (going back past the current 18 month window excludes the 10th Anniversary Set, Gehenna and KMW. In a month or two, you can add Legacies to the list). Shouldn't a system ideally suited for predicting player skill take this into account? If not with a sliding time window on, how else do you accomplish this? De-activating a rating after a certain timeframe IMO doesn't cut it, either... It would seem an ELO system which didn't age the results of the participants would poorly predict performance for a person who chose not to participate in VTES for a while, since upon his return that person would be relatively unfamiliar with newer cards being played, and would suffer in performance as a result. DaveZ Atom Weaver

"James Coupe" <ja...@zephyr.org.uk> wrote in message news:NRnAiGin...@gratiano.zephyr.org.uk... > In message <Et8Pe.70648$DW1.47979@fed1read06>, Frederick Scott > <nos...@no.spam.dot.com> writes: >>But no one said he had only 12 tournaments in the whole of his careeer - >>only 12 in the last 18 months. To conjecture that these are the only >>12, you'd have to assume a third place finish in his first CQ was his >>first tournament ever. Does that seem very likely to you? > > If you're going to argue that we should maintain everyone's ratings > forever, you have two issues: > > 1) Impracticality. Robert Goudie has pointed out just how difficult it > is to continually recalculate ratings when errors are found. > (Duplicate membership numbers, or whatever.) > > 2) That the game changes. *claps* I didn't receive this until after I posted my own take on it. AGREED!!! Chess is popular because its been practically the same damned game for more than 3000 years. VTES doesn't hold still for ten months... A good predictor of player skill _must_ take this key aspect of the game into account. DaveZ Atom Weaver

I wrote: > Maybe. I'll go look at the numbers. I think it is safe to assume that a > ratio of VPs/Games is a reasonable indicator of some kind of "skill" > measurement. I'll crunch some numbers and see what comes up. So I went and crunched out the top 27 players in the world (I stopped at 27 'cause, well, I got bored and figured it was a pretty good spread :-). What I was looking at was total VP's gained in constructed games divided by total number of constructed games played, figuring that if someone had a rating by "convnetional" means, they'd have a reasonably high ratio of VP/games (it turns out that about 1.5 VP per game is a conventional good score) and if someone had a high rating via the "Tatu Factor", they'd have a noticably low ratio of VP/game. The numbers that came up tend (at least in my eyes) to support my claim that the "Tatu factor" doesn't have that much impact on the system as a whole (it has an impact on an individual rating, sure, but there aren't that many people for whom it has an impact, making them acceptibel outliers, if you will). Rank Name VP/Game ratio Total Constructed Games Played 1. Ruben Ramos 1.93 99 2. Ben Peal 1.63 197 3. Van Ruben 1.52 94 4. Stefan Ferrenci 1.82 74 5. David Armaing 1.74 57 6. Hugh Angsensing 2.10 87 7. Matt Morgan 1.61 85 8. Jay Kristoff 1.51 181 9. Martin Weinmayer 1.70 127 10. Damnas 1.57 118 11. Francois Morand 1.80 134 12. Erik Torstensson 2.03 170 13. Itmar Gonzales 1.40 124 14. Roberto Rueda 1.62 112 15. Kamel Sensei 1.73 160 16. Stephane Lavrut 1.88 99 17. David Tatu 0.87 223 18. Israel Barbero 1.55 116 19. Pierre Brouille 1.52 111 20. Benoit Oliveri 1.54 63 21. Karol Magda 0.75 76 22. Frenc Vasadi 1.50 80 23. Ivan Santamaria 1.41 137 24. Remy Auclair 1.57 71 25. David Fraile 1.67 124 26. David Gimenez 1.40 70 27. Trey Morita 1.41 120 Average VP/Game ratio: 1.58 Avergae number of games played: 115 (my apologies for mangling people's names--I was copying my own sketchy handwriting...) So now we have some numbers. Keep in mind that the specific rank is determined by the best 8 games, and bigger games are weighted higher than smaller games, so it isn't unusual that, like, Van Ruben (1.52) is higher than Stefan (1.82). The average VP/Game ratio is 1.58. Of the 27 players used as a sample, most people hover around the 1.58 mark. There are, what, 2 people over 2.0 and 2 people under 1.0. Looking specifically at Mr. Tatu (0.87), for whom the factor is named, he clearly is an outlier--his VP/Games ratio is about 50% of the average, and his games played are about 200% the average. He is clearly benefiting from the system, in terms of his ranking (and again, this is likely 'cause David likes taking risks in deck choice in competition, not 'cause he is a questionable player). His large number of games totally makes up for his low aquisition of VPs over the long run. The other player with a low VP/Game ratio is Karol Magda (0.75), but he hasn't actually played all that many games--at 76 games, he is significantly below the average number of games played (66%) yet he has a high rating--he isn't gaining from the Tatu factor, he is scoring high for other not readily apparent factors--maybe he has done very well in a small number of prestige events (maybe he only plays in Qualifiers and Championships; maybe he has a tendancy to either win the tournament or score nothing--the VP/Game ratio doesn't take into account the weight of winning/placing in events). On the high end, Hugh Ansensing (2.10) also hasn't played all that many games (75% of average), but scores a lot of VPs in his games. Erik Torssensten (2.03) has played a lot of games (147% of average) and also scores a lot of VPs per game. Both of these are outliers, but reflective of strong play rather than middling play over a lot of games. What does this all say? The Tatu factor certainly can have a significant impact on the players for whom it impacts (in the top 27, that is, significantly, only Mr. Tatu himself), but overall, it doesn't come up that much--the other player under 1.0 isn't benefiting from the Tatu factor. Maybe I'll crunch the numbers of the next 23 players or whatever later. [ quoted text not captured ]

And here is part two of the top 50 players: 28. Antero Lappanen 1.98 83 29. Mikko Raimi 2.59 26 30. Miguel Pascual 1.81 74 30. Charles Leichausseur 1.36 83 31. David Quinonero 1.40 81 32. Anthony Coleman 1.59 59 33. Pierre Tran Van 1.34 99 34. Ville Kilpi 1.66 51 35. Marc Desaulnighy 1.70 41 36. Matej Lenareth 1.85 68 37. Robyn Tatu 1.31 221 38. Andrew Daley 1.80 92 39. Antione Franquinne 1.26 76 39. Miquel Ramos 1.53 68 40. Attila Sipos 1.34 123 41. Dave Pennington 1.32 81 42. Elol Ongun 1.47 96 43. Weverton Guilmero 1.73 67 44. Peter Raphail 1.71 88 45. Chris Meland 1.46 88 45. Dieter Ahrwellier 0.45 97 46. Mark Loughman 1.25 115 47. Brad Cashdollar 1.12 80 Average of these 23 VP/Game: 1.52 Average of these 23 total games: 85 Average of total top 50 VP/Game: 1.55 Average of top 50 total games: 101 Right--so what do we have here? Looking at the arbitrarily seperated low 23 as opposed to high 27 (about half and half), the VP/Game ratios are much more varried, but the average games played are much lower than with the top 27 (85 vs 115). Indicating that the more games you play, the more your score averages out (a lead article from the journal of "Duh":-). Looking at the bottom 23, again, most of the people, still, hover around the average of 1.5 something--the top 27 are all a bit higher, the lower 23 are all a bit lower, on average (compare the averages :-). If we look for the outliers in this group (over 2.0 and under 1.0), we get 1 guy with an incredibly high 2.59 (Mikko), but then he also has the lowest total of games out of everyone in the top 50 (26, or 25% of the average)--he likely has played not many events but has done really well in the ones he has played--as he plays more games, he will likely average out more. We also have 1 guy under 1.0--Dieter at the astonishing 0.45 across 97 games. Still, he is playing fewer games than total average (97 of 101 for the top 50; a little more than average from the group of 23)--so he is hardly scoring on the Tatu factor. He is probably in the same boat as Karol Magda from the upper 27--not that good performance across the board, but he is likely doing well in prestige events. Notably, Robyn Tatu is significantly *not* benefiting from the Tatu Factor--yeah, she is playing lots of games (the second most on the list of 50; more than twice the total 50 average), but she has a reaosnable VP/Game ratio of 1.31--lower than average, but not so much that she is significantly gaining from the Tatu Factor. Which is funny, what with her, ya know, being a Tatu... [ quoted text not captured ]

On Thu, 25 Aug 2005, Peter D Bakija wrote: > The numbers that came up tend (at least in my eyes) to support my claim that > the "Tatu factor" doesn't have that much impact on the system as a whole (it > has an impact on an individual rating, sure, but there aren't that many > people for whom it has an impact, making them acceptibel outliers, if you > will). <snip data and stuff> Might be more worthwhile to look at game wins rather than VPs since getting lots of game wins is more important than getting lots of VPs (although the latter certainly contribute to the fomer), but the numbers will probably be at least somewhat similar. For my money, there is no such thing as a "Tatu Factor." If there were one single example of this phenomenon other than David Tatu, I might buy it. As for David, he doesn't have a low percentage of VPs/wins because he's not a good player. He has a low percentage because he's often playing some kind of untried, questionable or even bad tech. If he played his best decks every tournament, his percentage and rating would be a lot higher. I imagine he plays all those goofy decks because he plays in so many tournaments and it's more exciting to win with something different or innovative than it is to win with the same old deck he's already won a few tournaments with, but I don't know for certain, not having asked him. I know you weren't trying to prove that David is ranked highly because of a scattershot approach to tournament play, Peter. I just thought this would be a good opportunity to attempt to dispell what I believe to be a myth. In general, your analysis shows that many of the top players score around the same number of VPs per game, which is something we should expect. Matt Morgan

"James Coupe" <ja...@zephyr.org.uk> wrote in message news:NRnAiGin...@gratiano.zephyr.org.uk... > In message <Et8Pe.70648$DW1.47979@fed1read06>, Frederick Scott > <nos...@no.spam.dot.com> writes: >>But no one said he had only 12 tournaments in the whole of his career - >>only 12 in the last 18 months. To conjecture that these are the only >>12, you'd have to assume a third place finish in his first CQ was his >>first tournament ever. Does that seem very likely to you? > > If you're going to argue that we should maintain everyone's ratings > forever, you have two issues: > > 1) Impracticality. Robert Goudie has pointed out just how difficult it > is to continually recalculate ratings when errors are found. > (Duplicate membership numbers, or whatever.) Actually, I don't think it's as difficult as all that. You only have to recalculate from when the error started making a difference. It seems unlikely that an error which is important enough to recalculate years worth of results will get unearthed and be deemed important enough to do so. (If we find out Fred Scott and Scott Fred are the same guy and the latter was a VEKN number used to record a single tournament result from six years back, would we really care? Would we really bother to recalculate everything for that?) But ultimately, if deemed necessary, it's completely doable - just a major P-I-T-A. And, as time goes by, disk drives get bigger, computers get faster, and the whole thing becomes even less of a challenge. The real challenge that prevents it right now is programming the original system, which would have to be sophisticated enough to be able to slide alterations in and recalculate from an arbitrary point. The much more common situation than duplicate numbers would be late tournament results. But I'd guess the number crunching power shouldn't be that much of an issue except when going back a long ways, and even then it's only a matter of dedicating the computer for long enough periods. > 2) That the game changes. It is entirely possible for someone to have > been plugging away at a given strategy, but only to finally push > it over the edge on the release of a given set. Say, you want > to play Setite Corruption strategies. Results prior to the > Final Nights (or KMW, or...) set might have you underperforming, > because you're plugging away at a strategy that doesn't work so > well. Then, oh look, the perfect card turns up for you and > bang, you're away. I'm not sure I see the issue. How is this any different from just getting better at the game? Sure - player ability changes over time, for better or for worse. Usually not so quickly that the player's rating shouldn't be able to follow but, as everyone seems to be fond of saying in this debate, no system is perfect. > How good someone was at playing in the environment three years ago is > not the same as how good they are now. Bear in mind, for instance, that > many political players used table-seating changing votes as a matter of > course and got many, many VPs they wouldn't get now as a result. No, but I guess I disagree that player abilities change so capriciously and so much as you seem to think. No one I've ever met plays exclusively one type of deck even if most players gravitate to certain types of decks. I do think results from 3 years ago are reasonable as data. Don't forget that as each new result comes, preceding results slide downward or "fade" in terms of importance. It may not be perfect data to use but in most cases it's much better than flat expiring it at the 18 month mark - and horribly biasing your system in favor of players who play more in the process. > And, I've also met several good players who simply don't like > tournaments - they don't like the places they're held, they don't like > giving up all day to it, or whatever. But some of them are very, very > good players. Lady Legbiter would be one example - very good player, > doesn't like tournaments. > > What would you rating system do when players like that choose to play? What can you do? Lady Legbiter is a perfect example of why the current system is horrible: she wouldn't have many points not because she isn't good but because she doesn't have many recent results. Use her older results, which may not be perfect but ought to be at least reasonable. Fred

Frederick Scott wrote: > "James Coupe" <ja...@zephyr.org.uk> wrote in message > news:NRnAiGin...@gratiano.zephyr.org.uk... > > In message <Et8Pe.70648$DW1.47979@fed1read06>, Frederick Scott > > <nos...@no.spam.dot.com> writes: > >>But no one said he had only 12 tournaments in the whole of his career - > >>only 12 in the last 18 months. To conjecture that these are the only > >>12, you'd have to assume a third place finish in his first CQ was his > >>first tournament ever. Does that seem very likely to you? > > > > If you're going to argue that we should maintain everyone's ratings > > forever, you have two issues: > > > > 1) Impracticality. Robert Goudie has pointed out just how difficult it > > is to continually recalculate ratings when errors are found. > > (Duplicate membership numbers, or whatever.) > > Actually, I don't think it's as difficult as all that. You only have > to recalculate from when the error started making a difference. Yep. However, back when we used to do this, we didn't have a database that contained all results. Teh ratings coordinator would have to manually work through all of the Archons. I'm sure, for example, Magic has a database that is capable of making this work. I doubt VTES could justify the expense of creation and maintenance of this setup, however. -Robert

"Frederick Scott" <nos...@no.spam.dot.com> wrote in message news:VxoPe.70942$DW1.238@fed1read06... > "James Coupe" <ja...@zephyr.org.uk> wrote in message > news:NRnAiGin...@gratiano.zephyr.org.uk... >> In message <Et8Pe.70648$DW1.47979@fed1read06>, Frederick Scott >> <nos...@no.spam.dot.com> writes: >>>But no one said he had only 12 tournaments in the whole of his career - >>>only 12 in the last 18 months. To conjecture that these are the only >>>12, you'd have to assume a third place finish in his first CQ was his >>>first tournament ever. Does that seem very likely to you? >> >> If you're going to argue that we should maintain everyone's ratings >> forever, you have two issues: >> snip #1 > >> 2) That the game changes. It is entirely possible for someone to have >> been plugging away at a given strategy, but only to finally push >> it over the edge on the release of a given set. Say, you want >> to play Setite Corruption strategies. Results prior to the >> Final Nights (or KMW, or...) set might have you underperforming, >> because you're plugging away at a strategy that doesn't work so >> well. Then, oh look, the perfect card turns up for you and >> bang, you're away. > > I'm not sure I see the issue. How is this any different from just > getting better at the game? Sure - player ability changes over > time, for better or for worse. Usually not so quickly that the player's > rating shouldn't be able to follow but, as everyone seems to be fond of > saying in this debate, no system is perfect. > James isn't talking about player ability changing, he's talking about the game (rules and game components) changing... The player who is inexperienced in the current meta-game (including most recent sets) with prior good performance will have poorer predictability of future performance in a systme which doesn't devalue older results on some a time scale. >> How good someone was at playing in the environment three years ago is >> not the same as how good they are now. Bear in mind, for instance, that >> many political players used table-seating changing votes as a matter of >> course and got many, many VPs they wouldn't get now as a result. > > No, but I guess I disagree that player abilities change so capriciously > and so much as you seem to think. No one I've ever met plays > exclusively one type of deck even if most players gravitate to certain > types of decks. I do think results from 3 years ago are reasonable > as data. You do? Three years ago (August 25 2002), there was no Anarchs, no Black Hand, no Gehenna, no 10th Anniversary, no Kindred Most Wanted, and Camarilla Edition was all of six days old (not valid for tournament play for another 24 days)... How do you consider that environment to be _anything_ like the VTES of today? I'll use myself as an example here. Three years ago, I could easily hold my own with any player from Atlanta, Columbia, Raleigh-Durham, or Boston (the groups I most commonly played against). I was winning drafts in Atlanta, and making finals in constructed events consistently, against some of the best competition on the East Coast. In the intervening time, a relocation and two kids has _severely_ cut down my play time, and I know that I'm no where near as polished a player as I once was... A rating system without an expiration date would show me as a far greater threat in the next tournament down the road than is at all warranted. I recall having to regularly ask for newer cards to be read aloud at TotalCon (Feb 05), and felt that my weak(er) performance there (8th or 12th in the New England qualifier, IIRC) had a lot to do with rustiness and a lack of familiarity with newer cards. > Don't forget that as each new result comes, preceding results > slide downward or "fade" in terms of importance. It may not be > perfect data to use but in most cases it's much better than flat > expiring it at the 18 month mark - and horribly biasing your system > in favor of players who play more in the process. > You have yet to show (other than the sole example of the Tatu Factor) the extent to which the current system is 'horribly biased'. At least Bakija is willing to crunch a few numbers... _You're_ the one with the issue with the current system (and have plenty of time to post about it, again and again and again), but you can't take the time to dig up the stats to back your assertions? May I advise less posting of the same old assertions (they're all archived in triplicate for anyone who wants to read them), and more time with the actual results showing where they are bad. We won't miss your fourteenth post about how bad the current system is, honest. You can take that time and use it to show us why... Show someone an incidence rate of badly mis-rated players which shows an issue, and you might start convincing people there's a problem (you'll convince a lot more people than you do by re-posting your SOS on the topic, at the very least..) DaveZ Atom Weaver

"Matthew T. Morgan" <far...@io.com> wrote in message news:2005082512...@eris.io.com... [ quoted text not captured ] ...which begs the question, would a more rigorous system (one which uses every tournament result generated for all time) discourage tournament innovation? IMO, it would. Also IMO, that in turn would ruin one of the reasons for participating in tournaments in the first place (seeing the interesting, innovative or just plain crazy tech others attempt)... It would also discourage participation in tournaments where the meta-game was unknown (traveling participants) if those travelers were at all concerned about their rating. Better the meta-game (devil) you know... I'd rather keep the tournament rating system as one which encourages innovation, cross participation and allows some amount of risk-taking with deck design, than to have one wherein the focus shifts to flat-out performance. It seems the former is the better for a 'cult' game like VTES, and works to keep things more interesting. DaveZ AW

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF2F631D.215FB%pd...@lightlink.com... > david.che...@gmail.com wrote: > >> There's no practical reason why the final can't be raised to 2.5 >> hours. >> The game will not just expand like a gas to consume whatever time you >> throw at it. People just believe that because they'll argue against >> any change at whatever cost, generally speaking. The status quo is >> god. > > I'd certainly be in favor of upping, at the very least, the length of > finals > for, like, national championships or whatever to 3 hours. But then, > there > are those that would argue that, in fact, games will expand like gas, > and > that 3 hour finals would time out just as often as 2 hour finals. It > would > just take longer to time out. I'm not necessarily one of those > people--I > rarely see games time out in competetive play, but then as I have > pointed > out elsewhere, I'm what I like to call a "load bearing" player, in > that I > either win or die trying, which speeds the whole game up for everyone, > but > when I do see games time out, they rarely are games that are almost > over but > just run out of time, they are usually games that have hit a stasis > wall, > and an extra hour would generally result in the game timing out in an > extra > hour. I believe you and David are right that games *would* end more often if the time limit were a bit longer (for the finals at least). There would be some gaseous expansion, but sure, probably not enough to totally cancel out the benefit. The problems I envision are more in the neighborhood of: 1. It makes the finals a little more of a different game than the preliminary rounds - decks that do well in 2-hour games may not do as well in 2.5 hour games, and vice versa. 2. The finals are already (in my experience) sometimes the worst game of VTES you play that day: I've just finished playing six hours of VTES and my brain is fried. Now I play the one that counts for another two hours. Making that even longer isn't going to make me play better. :-) 3. A fair number of people already think VTES tournaments take too long and have trouble fitting them into their schedules. Lengthening them even further does add to that issue. > In any case--congrats to Peter X of Michigan. And special props to > ex-Ithaca-home-team-member Joshy boy! Thanks! Everything I know about VTES, I learned in Ithaca. :-) (Remember those old Ventrue decks with too much combat defense I used to play, and then I'd kill you when my vampires refused to die? Ah, the good old days...) Josh skin of steel, obedience, majesty?

"David Zopf" <david...@snetx.net> wrote in message news:pAjPe.3265$L77....@newssvr19.news.prodigy.com... > "Frederick Scott" <nos...@no.spam.dot.com> wrote in message > news:kE7Pe.70483$DW1.3840@fed1read06... >> >> "James Coupe" <ja...@zephyr.org.uk> wrote in message >>> If the best player IN THE WORLD stopped playing tournaments for a while >>> and (therefore) didn't generate enough useful data, but then came back >>> and won a major tournament, would you point and say "Hah, your rating >>> system doesn't work"? >> >> You're assuming that if someone stops playing, you don't use past >> data. Why are you assuming that? >> >> The current system has to do something like that because it insists >> on using participation as part of its rating and must therefore slide >> the time period from which it considers results along with the date. >> (Past 18 months, in this case.) If you don't construct your system >> that way, you're under no obligation to throw out old results just >> because they're old. > > You've raised this point before(ad nauseam)... you seem to be neglecting > an aspect of VTES that to me seems both obvious and critical. Unlike chess, > bridge, or any of the other rated games used as a comparitor in this thread, > VTES (both its rules and its play components) changes over time. It could > just as easily be said that the rating system slides the time period (very > slowly) because the game of VTES today is significantly different from the > game of VTES played 18 months ago. You and James are only now bringing up this "obvious and critical" point, which doesn't strike me as either. Sure, the metagame changes somewhat over time but I don't think that means player skill changes that much just because the metagame changes. Some players may be more or less comfortable in different metagames, but so what? Jyhad is different from Chess in lots of ways, probably much more critically different in how many types of luck are present and how much they influence the game. But let's say it _does_ affect certain players in discernibly positive or negative ways. What are you going to do? You can throw up your hands and declare the game unratable - and thus all opinions of the quantity of skill you or other players possess must be conceded as utter bullshit. (In no patterns can be identified through numerical analysis of results, then what possible meaning could subjective opinions hold?) Or, I suppose you expire results chronologically and thus Matt's and James's scenario - where Peter Charnley's rating whipsaws up and down because he basically doesn't have an appropriate number of results in the most recent year or two. That, to me, is little improvement although you could certainly eliminate the bias towards more participation by making other changes. In the end, the "sliding metagame" factor just doesn't worry me or I'd have never posted anything about Peter Charnley. Obviously, I think he must be better than his rating shows to make the finals of the NAC twice in two years in a row - both in last year's metagame and this year's. A lot of the guys we acknowledge as good seem to be able to do that. > It would seem an ELO system which didn't age the results of the > participants would poorly predict performance for a person who chose not to > participate in VTES for a while, since upon his return that person would be > relatively unfamiliar with newer cards being played, and would suffer in > performance as a result. I agree, ELO would be vulnerable to that. Hell, Chess is probably vulnerable to that. In both cases, you get out of practice and your rating as you reenter will not properly reflect your skill at that moment. I'll grant, Jyhad has more issues because of the changing metagame and (to add to your complaint) because the person might not own critical cards that have become available. Still, in the end, I consider these to be minor issues and think the results would be far more accurate overall for purposes of reflecting skill than the current system. What your complaints mainly do is just bring into question how accurate any rating system for Jyhad could ever be. Fred

Matthew T. Morgan wrote: > Might be more worthwhile to look at game wins rather than VPs since > getting lots of game wins is more important than getting lots of VPs > (although the latter certainly contribute to the fomer), but the numbers > will probably be at least somewhat similar. Hmm. That might be a good idea. I'll try that one next... > For my money, there is no such thing as a "Tatu Factor." If there were > one single example of this phenomenon other than David Tatu, I might buy > it. As for David, he doesn't have a low percentage of VPs/wins because > he's not a good player. He has a low percentage because he's often > playing some kind of untried, questionable or even bad tech. If he played > his best decks every tournament, his percentage and rating would be a lot > higher. I imagine he plays all those goofy decks because he plays in so > many tournaments and it's more exciting to win with something different or > innovative than it is to win with the same old deck he's already won a few > tournaments with, but I don't know for certain, not having asked him. Oh--I'm totally with you here (in fact, I think I wrote the exact same paragraph about David, verbatim, in my report somewhere :-) But for Fred's sake, who is convinced that "The Tatu Factor" has a huge impact on the validity of the rating system, I was clearly illustrating that it really only had an effect on, well, David. > I know you weren't trying to prove that David is ranked highly because of > a scattershot approach to tournament play, Peter. I just thought this > would be a good opportunity to attempt to dispell what I believe to be a > myth. In general, your analysis shows that many of the top players score > around the same number of VPs per game, which is something we should > expect. Yep. The top 50 players are all hovering around the 1.5 VPs per game. I'll have to spend more time and see when the ratings even out at around 1 VP per game, which is what I'd expect to be the absolute average level of performance. [ quoted text not captured ]

David Zopf wrote: > ...which begs the question, would a more rigorous system (one which uses > every tournament result generated for all time) discourage tournament > innovation? IMO, it would. I think it certainly would--which is why I think the "only your 8 best games" is a concrete *benefit* of the system--it doesnot punish you for playing wacky decks occasionally, which means more varried tournaments and more interesting play environments. If your rating kept track of *every* tournament you played (ya know, assuming you care about your rating), no one would ever play anything other than the cannon of tournament winning decks (ya know, S+B, weenie DOM, Ventrue Law Firm, whatever). As the system currently works, you can play crazy decks without harming your rating, assuming you do well in at least 8 events. [ quoted text not captured ]

Joshua Duffin wrote: > 1. It makes the finals a little more of a different game than the > preliminary rounds - decks that do well in 2-hour games may not do as > well in 2.5 hour games, and vice versa. True. > > 2. The finals are already (in my experience) sometimes the worst game of > VTES you play that day: I've just finished playing six hours of VTES and > my brain is fried. Now I play the one that counts for another two > hours. Making that even longer isn't going to make me play better. :-) Also true. > > 3. A fair number of people already think VTES tournaments take too long > and have trouble fitting them into their schedules. Lengthening them > even further does add to that issue. Very, significantly true. > Thanks! Everything I know about VTES, I learned in Ithaca. :-) > (Remember those old Ventrue decks with too much combat defense I used to > play, and then I'd kill you when my vampires refused to die? Ah, the > good old days...) Man. It was like you were magic--always throwing the Rock to my Scisors. Jason keeps doing that these days. [ quoted text not captured ]

Peter D Bakija wrote: > Joshua Duffin wrote: > > Thanks! Everything I know about VTES, I learned in Ithaca. :-) Does it count if I say that everything I know about VTES, I learned from the Ithaca players? :P > Man. It was like you were magic--always throwing the Rock to my Scisors. > Jason keeps doing that these days. Good old rock. Nothing beats rock! On a related note, I think that Josh is right in that VTES tournaments already take up all day. Adding another half an hour isn't a big deal to those of us that already play them with some frequency, but making them longer isn't going to win over anyone already on the edge. Xian it's true, though...

I wrote: > Hmm. That might be a good idea. I'll try that one next... And here are the stats with GW/Games: Rank Player VP/Games Total Games GW/Games 1. Ruben Ramos 1.93 99 .47 2. Ben Peal 1.63 197 .37 3. Van Ruben 1.52 94 .34 4. Stefan Ferrenci 1.82 74 .43 5. David Armaing 1.74 57 .45 6. Hugh Angsensing 2.10 87 .49 7. Matt Morgan 1.61 85 .35 8. Jay Kristoff 1.51 181 .30 9. Martin Weinmayer 1.70 127 .40 10. Damnas 1.57 118 .32 11. Francois Morand 1.80 134 .44 12. Erik Torstensson 2.03 170 .47 13. Itmar Gonzales 1.40 124 .32 14. Roberto Rueda 1.62 112 .36 15. Kamel Sensei 1.73 160 .37 16. Stephane Lavrut 1.88 99 .49 17. David Tatu 0.87 223 .17 18. Israel Barbero 1.55 116 .31 19. Pierre Brouille 1.52 111 .34 20. Benoit Oliveri 1.54 63 .38 21. Karol Magda 0.75 76 .47 22. Frenc Vasadi 1.50 80 .36 23. Ivan Santamaria 1.41 137 .29 24. Remy Auclair 1.57 71 .32 25. David Fraile 1.67 124 .39 26. David Gimenez 1.40 70 .34 27. Trey Morita 1.41 120 .31 28. Antero Lappanen 1.98 83 .44 29. Mikko Raimi 2.59 26 .61 30. Miguel Pascual 1.81 74 .43 30. Charles Leichausseur 1.36 83 .31 31. David Quinonero 1.40 81 .30 32. Anthony Coleman 1.59 59 .27 33. Pierre Tran Van 1.34 99 .28 34. Ville Kilpi 1.66 51 .39 35. Marc Desaulnighy 1.70 41 .39 36. Matej Lenareth 1.85 68 .39 37. Robyn Tatu 1.31 221 .24 38. Andrew Daley 1.80 92 .42 39. Antione Franquinne 1.26 76 .30 39. Miquel Ramos 1.53 68 .33 40. Attila Sipos 1.34 123 .27 41. Dave Pennington 1.32 81 .25 42. Elol Ongun 1.47 96 .29 43. Weverton Guilmero 1.73 67 .32 44. Peter Raphail 1.71 88 .35 45. Chris Meland 1.46 88 .29 45. Dieter Ahrwellier 0.45 97 .20 46. Mark Loughman 1.25 115 .22 47. Brad Cashdollar 1.12 80 .22 Average of total top 50 VP/Game: 1.55 Average of top 50 total games: 101 Average of top 50 GW/Game: 0.35 Right. So now what does this tell us? Looking over the list, emphasizing Game Wins per game rather than VP per game, the list averages out even more. There are now only two significant outliers--one player with a score over 0.50 (Miiko Raimi at 0.61, who is also significantly the player with the fewest games on the list) and only one player with a score under 0.20 (David Tatu with a 0.17). Everyone else falls between 0.20 and 0.50--if someone were to make a nice scatter graph of the results, it seels likely that it would look quite average, with Miiko significantly above the line and Tatu significantly below the line. The other noticable outliers from the VP/Game analysis cease to be outliers in the GW/Game analysis (Karol's low .75 VP/Game becomes a very respectable 0.47 GW/Game; Hugh Ansengsing's high 2.10 VP/Game becomes a similarly acceptable .049 GW/Game). So I would tend to agree with Matt Morgan's analysis of the situation--The "Tatu Factor" tends to be completely insignificant in the grand scheme. Of the top 50 players in the world, the only player for whom the Tatu Factor is significant is, appropriately, David Tatu--even Robyn Tatu (the player with the second most games in the system, and, ya know, also a Tatu) has a reasonable 0.24 GW/Game stat--below average, yeah, but 5 total players in the top 50 have a GW/Game of 0.20-0.25, and all of them are below rank 36--the average GW/Game of ranks 31-50 (i.e. bottom 20 of the top 50) is 0.30. [ quoted text not captured ]

Xian wrote: > On a related note, I think that Josh is right in that VTES tournaments > already take up all day. Adding another half an hour isn't a big deal > to those of us that already play them with some frequency, but making > them longer isn't going to win over anyone already on the edge. > Man. Those weak hearted fools... [ quoted text not captured ]

>>> I'm sorry, Peter. I'm saving myself for Wes. >> >> Man. Wes gets all the breaks. > > and if I remember right, also on the bubble just outside of > being in the finals of the Shadow Twin draft tournament on Saturday. Wes nearly won Shadow Twin Constructed on Friday.

[ quoted text not captured ] *shrug* two things held me back. 1) I was hoping you'd just stop posting the same old shit. 2) My ability to post in the daytime is restricted by my job responsibilities. It seems others aren't under such heinous restrictions, so i generally leave most debate to them... > Sure, the metagame changes somewhat > over time but I don't think that means player skill changes that much just > because the metagame changes. Some players may be more or less comfortable > in different metagames, but so what? Jyhad is different from Chess in lots > of ways, probably much more critically different in how many types of luck > are present and how much they influence the game. > I'm not talking about meta-game (defining that as the "local play group variation of deck selection, tactics and style), I'm talking about the game itself, Fred. The cards and the rules. They change over time. You don't think so? In the last three years, there are 115 new Came Ed cards 122 new Anarch cards 138 new Black Hand cards 150 new Gehenna cards 10 new 10th Anniversary cards 150 new Kindred Most Wanted cards 11 promo cards ...for a total of 696 new cards, out of 2035 total cards in the game. Thats a hair more than 30% of all the cards in the game, Fred! Pick any 18 month timeframe which cuts out two expansions, and you're still talking about roughly 15% of the cards in the game... I haven't even yet tried to count the cards which have gotten re-writes which effectively alter their in-game function. > But let's say it _does_ affect certain players in discernibly positive or > negative ways. What are you going to do? You can throw up your hands > and declare the game unratable - and thus all opinions of the quantity > of skill you or other players possess must be conceded as utter bullshit. ...or you can have ratings which are old simply matter less (or at some point not at all), which will give you a lower rating, which in turn will appropriately reflect lack of experience with VTES as it is _today_. > (In no patterns can be identified through numerical analysis of results, > then what possible meaning could subjective opinions hold?) Or, I suppose > you expire results chronologically and thus Matt's and James's scenario - > where Peter Charnley's rating whipsaws up and down because he basically > doesn't have an appropriate number of results in the most recent year or > two. That, to me, is little improvement although you could certainly > eliminate the bias towards more participation by making other changes. > > In the end, the "sliding metagame" factor just doesn't worry me or I'd > have never posted anything about Peter Charnley. Obviously, I think > he must be better than his rating shows to make the finals of the NAC > twice in two years in a row - both in last year's metagame and this > year's. A lot of the guys we acknowledge as good seem to be able to > do that. > Peter's a mediocre example of the effect I'm driving at, since he has at least been participating in some capacity in the intervening time between NACs... You've got enough of something to hang an inkling of a rating off of (and I'm with Peter, 27th going in with the results he had prior to NAC'05 is absolutely reasonable, for the results he had garnered up to that point). > > It would seem an ELO system which didn't age the results of the > > participants would poorly predict performance for a person who chose not to > > participate in VTES for a while, since upon his return that person would be > > relatively unfamiliar with newer cards being played, and would suffer in > > performance as a result. > > I agree, ELO would be vulnerable to that. Hell, Chess is probably > vulnerable to that. In both cases, you get out of practice and your > rating as you reenter will not properly reflect your skill at that > moment. I'll grant, Jyhad has more issues because of the changing > metagame and (to add to your complaint) because the person might not > own critical cards that have become available. Still, in the end, > I consider these to be minor issues and think the results would be > far more accurate overall for purposes of reflecting skill than the > current system. What your complaints mainly do is just bring into > question how accurate any rating system for Jyhad could ever be. > Again, I disagree. I don't think that 30% of all the cards ever printed is minor. There is a way to have the rating take this aspect of the game into account, and I think that the current system has a way of doing it. Whether it could be improved by changing the aging of results is certainly open for discussion. DaveZ Atom Weaver

And just 'cause I'm on fire, I went and ran the same numbers for the top 10 players in the North East of the US ('cause I know everyone on the list, including me...), and figured it was likely a good cross section of kindof middle players, nationally speaking (I'm number 4 in the NE US, but number, like, 67 world wide. So this is a group of people from the not top 50, mostly). Rank in NE US Player Games VP/Games GW/Games 1. Ben Peal 197 1.63 .37 2. Ben Swainbank 126 1.70 .41 3. Scott Gomes 127 1.09 .20 4. Peter Bakija 64 1.52 .36 5. Matt Flint 132 1.34 .27 6. Matt Hirsch 164 1.14 .20 7. Nick Watkins 89 1.46 .30 8. Jon Scherer 28 1.64 .32 9. Andy Kempton 73 1.13 .16 10. Lance Shoppe 90 1.13 .24 Top 10 NE US average games: 109 Top 10 NE US average VP/Games: 1.37 Top 10 NE US average GW/Games: .28 (Top 50 world wide averages are 101/1.55/0.35) Of these (ranging from Ben Peal, who is #2 worldwide and Lance Shoppe who is #170 worldwide), again, mostly pretty average, compared to the top 50. A lower average overall, but then lower worldwide rankings overall. Only one player below 0.20 GW/Game--Andy Kempton with 0.16 (below David Tatu's 0.17), but Andy also played fewer than the average number of games for the sample and the top 50, and also has a reasonable 1.13 VP/Game (i.e. he wins more than 1 VP per game he plays)--it looks like he gets a reasonable number of VPs in competition, but doesn't win games all that often. And I, as it turns out, am wildly average with virtually identical VP/Game and GW/Game as the top 50 worldwide players. Again, it looks like no one is really gaming the system here--no one is getting an unusually high rating due to playing an inordinate number of games to balance out sketchy play--Andy is certianly benefitting from playing a lot of games compared to his game wins (by scoring VPs in most of those games), but he is still playing a below average number of games. I'm going to look at low-mid rank players now... [ quoted text not captured ]

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF334BDA.21745%pd...@lightlink.com... >I wrote: > >> Maybe. I'll go look at the numbers. I think it is safe to assume that a >> ratio of VPs/Games is a reasonable indicator of some kind of "skill" >> measurement. I'll crunch some numbers and see what comes up. > > So I went and crunched out the top 27 players in the world (I stopped at 27 > 'cause, well, I got bored and figured it was a pretty good spread :-). What > I was looking at was total VP's gained in constructed games divided by total > number of constructed games played, figuring that if someone had a rating by > "convnetional" means, they'd have a reasonably high ratio of VP/games (it > turns out that about 1.5 VP per game is a conventional good score) and if > someone had a high rating via the "Tatu Factor", they'd have a noticably low > ratio of VP/game. > > The numbers that came up tend (at least in my eyes) to support my claim that > the "Tatu factor" doesn't have that much impact on the system as a whole (it > has an impact on an individual rating, sure, but there aren't that many > people for whom it has an impact, making them acceptibel outliers, if you > will). (One problem right off is that this says nothing of who plays against who, thus correcting for opponents' skill. But never mind that; I understand your analysis was meant to be crude. Even so...) Sorry, Peter. If you want to show this sort of correlation, you're coming up WAY short. It's not enough to demonstrate that the existing top-X list (whatever X is) seems to be a high number. You actually have no idea what those numbers mean. I agree, a higher than 1-per-game ration is obviously good, but how high should it be? What does a difference between two scores mean? Is a difference of 0.1 big or small? Who knows, without calculating every player's rating, from top to bottom, and then doing statistical analysis on it? The most obvious flaw (though not the only one) is the same one that I've pointed at all along: that good players with high VP/Games comparable to or better than these may be a ways down the list - where you can't see them just by calculating the Top 27. ... (ellided a lot of more or less random comments about VP/Game ratios of various top players) > What does this all say? The Tatu factor certainly can have a significant > impact on the players for whom it impacts (in the top 27, that is, > significantly, only Mr. Tatu himself), but overall, it doesn't come up that > much--the other player under 1.0 isn't benefiting from the Tatu factor. Huh? I'm completely confused what your criteria are for the Tatu factor "com(ing) up that much". The reason I brought David Tatu is that he doesn't seem to have played any better than Charnley by tournament placement yet he was rated much higher than Charnley. It was simply a convenient demonstration. It doesn't mean that one has to have a low VPs/Game average to have one's rating affected by the Tatu factor. EVERYONE is affected by the Tatu factor - except the very best (who have done near perfectly in 8 tournaments so far) and the very worst (who don't improve their ratings by playing more because they don't score at all). I'm pretty sure this is true because it because it has to be true. I don't think I have to convince you that first eight tournaments are important for player's to play. That's obvious. So from there: Tournament outcomes will vary. All things being equal, the ninth tournament has 8/9ths chance of being one of the player's top 8 tournaments. The average difference between a player who has played nine tournaments vs. one who has played eight tournaments depends on how much scores vary from one tournament to the next (the average deviation). This, in turn, will vary based on how good the player is and how large and long the tournaments are in which he tends to play. So all of this varies quite a bit and we'd have to do a lot of number crunching to understand how much a ninth tournament adds to the average player's score. But I think it should be pretty clear that with good players who have a fair chance of making any given tournament's final table, ON AVERAGE the ninth tournament will make a substantial difference in scores. Less so the tenth tournament, less so the eleventh and so forth. The return will reduce on a per-tournament basis as you go up on a curve which approaches zero. But it isn't clear to me that it reduces on a very steep curve. And notice that it actually has a _greater_ effect - not a lesser effect - for players who score more victory points per game. I don't know why you're assuming that a high number of VPs/game shows the Tatu factor is meaningless. To me, it just shows that all these guys have a very good chance of raising their rating - and thus their ranking - by simply attending more tournaments. Fred

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF339E9B.2176D%pd...@lightlink.com... > David Zopf wrote: >> ...which begs the question, would a more rigorous system (one which uses >> every tournament result generated for all time) discourage tournament >> innovation? IMO, it would. > > I think it certainly would--which is why I think the "only your 8 best > games" is a concrete *benefit* of the system--it doesnot punish you for > playing wacky decks occasionally, which means more varried tournaments and > more interesting play environments. Oh fer crying out loud. I think you all overestimate the effect of the rating system. Again, it's not for encouraging or discourging anything and I don't think it has that effect. Peter, do you recall when we had the original whacky ELO system in place?!? Did you ever even once worry about what your number was as you decided what deck to play? Do you know anyone else who did? I'd be shocked. The only "benefit" this system has in this sense is that it's so totally meaningless and whacked, that it's hard to care about it at all. If you want to the total benefit package, you should get rid of it altogether and replace it with nothing so rating systems can never affect anyone's thinking. But in fact, if people really do care about scoring well, they'll play a serious deck because that way they'd have a much better chance of cracking their lowest top 8 score. So how is this different than ELO? Fred, waltzing on past the silly and approaching the insane

Peter D Bakija wrote: > I'm going to look at low-mid rank players now... Looking at players 41-50 ranked in the US: Rank Player Games VP/Game GW/Game 41. Lance Shoppe 90 1.12 .24 42. Albert Lee 28 1.19 .21 43. Pat Lusk 43 1.16 .25 44. Tom Mickle 50 1.61 .36 45. Boris Zaretsky 57 1.44 .33 46. John Eno 23 1.19 .26 47. Mike Perlman 95 0.61 .08 48. Ira Fay 52 1.63 .36 49. Dave Wesener 20 1.20 .25 50. Donovan Brouwer 51 1.03 .17 Worldwide, we are spanning ranks 170-209. Nothing real surprising again--Mike Perlman seems to certainly be gaining from the system here, with a lot of games, not many VP's per game (0.61) and not many wins per game (0.08). Which brings us to 2 players out of about 70 who have ratings based on lots of games making up for middling performance. But only one in in the top 50. Strikes me as reasonable. Still. [ quoted text not captured ]

Frederick Scott wrote: > (One problem right off is that this says nothing of who plays against who, > thus correcting for opponents' skill. But never mind that; I understand > your analysis was meant to be crude. Even so...) That is because the rating system doesn't measure skill. It doesn't claim to measure skill. It only measures performance. I could make a little song about that if you'd like. > Sorry, Peter. If you want to show this sort of correlation, you're coming > up WAY short. It's not enough to demonstrate that the existing top-X list > (whatever X is) seems to be a high number. You actually have no idea what > those numbers mean. It doesn't matter what the numbers mean--what matters is that they indicate that players that are ranked highly are all ranked highly by, more or less, doing the same things. >I agree, a higher than 1-per-game ration is obviously > good, but how high should it be? What does a difference between two scores > mean? Is a difference of 0.1 big or small? Who knows, without calculating > every player's rating, from top to bottom, and then doing statistical > analysis on it? All I'm looking at here, really, is averages in performance--I'm not trying to come up with "skill" or whatever. The numbers show that most players tend to cluster around the same average performance level--the average of the top 50 players in the world is to get about 1.5 VP per game, winning about 1/3rd of the games they play, over about 115 games. And most players in the top 50 fit this model--the incidence of people having high ratings from playing lots of games and not doing so well is insignificant. Overall, the vast majority of players get their rankings by doing the same things that the other, similarly ranked players do. They play about the same numbers of games, they win about the same numbers of VPs per game, and they win about the same percentage of games. There is certainly *some* variation, but certainly among the top 50 players in the world, they are all doing pretty much the same thing to get that high ranking. > The most obvious flaw (though not the only one) is the same one that I've > pointed at all along: that good players with high VP/Games comparable to > or better than these may be a ways down the list - where you can't see > them just by calculating the Top 27. Sure. Because they are winning smaller events. Or not winning that many games. Players who do well at "prestige" events (the big ones) are going to be ranked higher than someone with similar VP/Game and GW/Game stats. The rating is not made by you VP/Game or GW/Game, but by the quality of the VPs and Games you win, as abstracted, reasonably, by virtue of the size and quality of the event--a VP won in an 8 person local tournament is worth less, in terms of ranking, than a VP won in a 60 person Qualifier. Which is fine. > Huh? I'm completely confused what your criteria are for the Tatu factor > "com(ing) up that much". It is insignificant *in terms of the total field of players*. I'm kind of confused as to how you can still not understand this point. Again, playing lots of games to make up for lackluster performance certainly can make a difference in a single person's rating (although apparently, it doesn't actually that much). But in terms of players who have a hig score based on many games disguising low scores overall, there is a total of 1 in the top 50 (Mr. Tatu). Everyone else in the top 50 got to the top 50 by doing the same thing--playing about the same number of games, winning about the same number of VPs per game, and winning about the same percentage of games. The difference at that point comes down to, more than anything, the "quality" of the points won--were the VPs and Game Wins from big events with high levels of competition or were they VPs and GWs from small, local events. Playing more big events pushes you up where playing small local events keeps you lower. > The reason I brought David Tatu is that he doesn't > seem to have played any better than Charnley by tournament placement yet > he was rated much higher than Charnley. Yes. 'Cause David is the outlier. The only one. > It was simply a convenient > demonstration. It doesn't mean that one has to have a low VPs/Game > average to have one's rating affected by the Tatu factor. EVERYONE > is affected by the Tatu factor - except the very best (who have done > near perfectly in 8 tournaments so far) and the very worst (who don't > improve their ratings by playing more because they don't score at all). Everyone to some extent is, yes, saved by playing lots of games. But again, looking at the top 50, they all got in the top 50 by doing, more or less, the same stuff. > I don't know why you're assuming that a high number of VPs/game > shows the Tatu factor is meaningless. To me, it just shows that > all these guys have a very good chance of raising their rating - > and thus their ranking - by simply attending more tournaments. *Everyone* has a good chance of raising their ratings by attending more tournaments. That is part of the system. Which measures performance and not skill. This being said, everyone who has a high ranking does so by virtue of similar play habits and similar win patterns. The Tatu Factor is meaningless in the sense that it does not have a significant impact on ratings for some and not others--everyone benefits from playing more games, and yet everyone (in the top 50) is playing about the same number of games. The only person in the top 50 who has a high rating (i.e. they are in the top 50 worldwide) with significantly below average performance (by game) is Tatu. Everyone else is in the same boat. [ quoted text not captured ]

Frederick Scott wrote: > Oh fer crying out loud. I think you all overestimate the effect of the > rating system. Again, it's not for encouraging or discourging anything > and I don't think it has that effect. Most of the time, you'll recall that I qualify statements like that with a (assuming you care about these things). I got lazy there and forgot to type it in. But I think it is inherrent in the discussion. Assuming you care about ratings, the current system provides incentive to play wacy decks sometimes, as it doesn't hurt your rating. If there was a system that punished you for taking risks, and people were concerned about maintaining there ratings, they would not take risks. The current system, at a certain point (i.e. after you have 8 events under your belt), does not punish you for taking risks. > Fred, waltzing on past the silly and approaching the insane I'm doing my best. [ quoted text not captured ]

"David Zopf" <david...@snetx.net> wrote in message news:V2pPe.3339$u_6...@newssvr17.news.prodigy.com... > "Frederick Scott" <nos...@no.spam.dot.com> wrote in message news:VxoPe.70942$DW1.238@fed1read06... >> I'm not sure I see the issue. How is this any different from just >> getting better at the game? Sure - player ability changes over >> time, for better or for worse. Usually not so quickly that the player's >> rating shouldn't be able to follow but, as everyone seems to be fond of >> saying in this debate, no system is perfect. >> > James isn't talking about player ability changing, he's talking about the > game (rules and game components) changing... The player who is > inexperienced in the current meta-game (including most recent sets) with > prior good performance will have poorer predictability of future performance > in a systme which doesn't devalue older results on some a time scale. I just don't believe the difference is what you seem to think it is. >> I do think results from 3 years ago are reasonable as data. > > You do? Three years ago (August 25 2002), there was no Anarchs, no Black > Hand, no Gehenna, no 10th Anniversary, no Kindred Most Wanted, and Camarilla > Edition was all of six days old (not valid for tournament play for another > 24 days)... How do you consider that environment to be _anything_ like the > VTES of today? I'm not sure I'd describe it as being totally different but never mind that. The point is, just because environment has changed doesn't by itself mean that a player won't do approxmately as well in the new environment as the old one unless there's a specific reason he won't. For instance, one reason he might not is if he's only comfortable playing Thaumaturgy combat decks and Thaumaturgy combat has become much less viable. But I don't think most players are so narrow. > I'll use myself as an example here. Three years ago, I > could easily hold my own with any player from Atlanta, Columbia, > Raleigh-Durham, or Boston (the groups I most commonly played against). I > was winning drafts in Atlanta, and making finals in constructed events > consistently, against some of the best competition on the East Coast. In > the intervening time, a relocation and two kids has _severely_ cut down my > play time, and I know that I'm no where near as polished a player as I once > was... A rating system without an expiration date would show me as a far > greater threat in the next tournament down the road than is at all > warranted. OK, but you're talking about rust as being the specific reason. Even rust flakes off pretty quickly, though - allowing a player to get back to being about the kind of player he used to be. Are you understanding that one of the changes you'd have to make is to tune the ELO coefficients DOWN to the point that a few bad games or even a few bad tournaments won't move your score all that much? How long would you expect to spend getting back to your old level? I believe your abilities and experience, in the long run, is going to be more important than rust. But if you slow down and play on a much reduced level over an indefinite period of time, sure - your skill will truly go down because you don't play enough. It happens. Just like peoples' skill goes up with more play. I'd still rather have the possibility for out-of-date ratings than the significant and certain bias based on participation we now have by a long shot. > You have yet to show (other than the sole example of the Tatu Factor) the > extent to which the current system is 'horribly biased'. At least Bakija is > willing to crunch a few numbers... Bakija's numbers don't mean anything, except that he misunderstands how to think about the bias. I don't see how the fact that most of the people on the top 50 list happen to have good VP/game ratios (would was anyone expecting, anyway) shows non-correlation between number of tournaments and rating points. If you have a high VP/game ratio or not, you will still have a higher score if you attend 12 tournaments than if you attend 8, a higher still score if you attend 16, and a higher still score if you attend 20 - at least on average. The bias is that two equal skilled players, whatever their VP/game ratio, will not be equally rated on average if one has attended signficantly more tournaments than the other. THAT'S what the "Tatu factor" is. I have no idea what Peter thinks it is. And that is an obvious thing, requiring no statistics to back it up. > _You're_ the one with the issue with the > current system (and have plenty of time to post about it, again and again > and again), but you can't take the time to dig up the stats to back your > assertions? The only thing I could use statistics for is demonstrate what the magnitude of the bias is. And this would depend on various factors such as how many points do players score on average depending on how good they are and what size tournaments they attend and mainly, how much do the finishes vary from one another - which shows how much of an advantage attending additional tournaments over 8 would be. You can't do this solely by digging up statistics kept by VEKN. You'd also have to do a statistical analysis on the randomness of finishes varying by skill level (how would one define that?) and plug that into formulas you'd have to invent around using the top 8 out of X finishes. I have a Bachelor's degree in mathematics and I don't have a clue where to start. (Granting I didn't take many statistics courses.) Do you? Fred

> Nothing real surprising again--Mike Perlman seems to certainly be > gaining from the system here, with a lot of games, not many VP's per > game (0.61) and not many wins per game (0.08). Which brings us to 2 > players out of about 70 who have ratings based on lots of games making > up for middling performance. But only one in in the top 50. Strikes me > as reasonable. Still. Fred won't stop until you've done everyone. Forever. And I like chikken. Ankur

Frederick Scott wrote: > So all of this varies quite a bit and we'd have to do a lot of number > crunching to understand how much a ninth tournament adds to the > average player's score... The R^2 value is a measure of the correlation between two variables a value of 1 = strong correlation, a value of 0 = no correlation. Using the figures elsewhere of Peter's, if you calculate the R^2 value of the correlation between Ranking and Games Played it has a value of 0.09, for the top 50. That is, there is not a strong correlation between Games Played and Ranking, which would be expected if the Tatu factor was very significant. Additionally, if you calculate the R^2 for the correlation between Ranking and VP/Game, it has a value of 0.10. That is, Ranking does not strongly correlate to this measure of "ability", either. However, if you calculate the R^2 for the correlation between Ranking and Wins/Game, it has a value of 0.20. Again this is not a strong correlation, but nonetheless the Ranking represents those players who won, more than it represents those players who played lots. Therefore, a ranking correlates more to winning games than playing lots. This seems all good. I guess, if you had access to the statistics for this period (last 18 months), rather than career stats, the correlation would be more meaningful. -- * lehrbuch (lehr...@gmail.com)

Frederick Scott wrote: > Bakija's numbers don't mean anything, except that he misunderstands how to > think about the bias. I don't see how the fact that most of the people > on the top 50 list happen to have good VP/game ratios (would was anyone > expecting, anyway) shows non-correlation between number of tournaments > and rating points. Oh, for the love of punk rock. What my numbers indicate is that everyone in the top 50 are doing the same thing, more or less, to get to the top 50. That they have good VP/Game ratios is unimportant. That they all have *similar* VP/Game ratios means that they are all doing about the same stuff to get a high ranking. Meaning that they get there by virtue of similar means. Meaning that the number of games they play, relative to each other, is mostly unimportant. Yes. You get more rating points by going to more tournaments. But the people who have the highest rating points all are, more or less, playing the same number of tournaments. And the concept of "lots of tournaments can disguise shoddy performance and result in a high ranking anyway" is almost completely irrelevant to the system. [ quoted text not captured ]

Frederick Scott wrote: > You can't do this solely by digging up statistics kept by VEKN. > You'd also have to do a statistical analysis on the randomness of > finishes varying by skill level (how would one define that?) You can't define that. Which is why it is good that the system we have doesn't measure skill. It measures performance. It seems like the only person who wants it to do something that it doesn't (i.e. measure skill) is you. [ quoted text not captured ]

Ankur Gupta wrote: > Fred won't stop until you've done everyone. Forever. And I like chikken. He won't stop then, either. 'Cause he has the steely tenacity. If only we could meet again, only that time as allies... [ quoted text not captured ]

"Frederick Scott" <nos...@no.spam.dot.com> wrote in message news:00wPe.71383$DW1.17917@fed1read06... > "David Zopf" <david...@snetx.net> wrote in message > news:V2pPe.3339$u_6...@newssvr17.news.prodigy.com... >> "Frederick Scott" <nos...@no.spam.dot.com> wrote in message >> news:VxoPe.70942$DW1.238@fed1read06... >>> I'm not sure I see the issue. How is this any different from just >>> getting better at the game? Sure - player ability changes over >>> time, for better or for worse. Usually not so quickly that the player's >>> rating shouldn't be able to follow but, as everyone seems to be fond of >>> saying in this debate, no system is perfect. >>> >> James isn't talking about player ability changing, he's talking about the >> game (rules and game components) changing... The player who is >> inexperienced in the current meta-game (including most recent sets) with >> prior good performance will have poorer predictability of future >> performance >> in a systme which doesn't devalue older results on some a time scale. > > I just don't believe the difference is what you seem to think it is. > It is for me, sister. And I think its a logical to derive as >>> I do think results from 3 years ago are reasonable as data. >> >> You do? Three years ago (August 25 2002), there was no Anarchs, no Black >> Hand, no Gehenna, no 10th Anniversary, no Kindred Most Wanted, and >> Camarilla >> Edition was all of six days old (not valid for tournament play for >> another >> 24 days)... How do you consider that environment to be _anything_ like >> the >> VTES of today? > > I'm not sure I'd describe it as being totally different but never mind > that. > The point is, just because environment has changed doesn't by itself mean > that a player won't do approxmately as well in the new environment as the > old one unless there's a specific reason he won't. I _did_ specifically point out that I was tlaking about those with a lack of current-time experience (say someone slacks off playing for a year, or doesn't buy in to a particular set, etc.), so I'll go along with the fact that there must be an additional reason besides change itself. If a player keeps buying and playing through the change, he'll be as well off as he ever was. But honestly, how many among us in the past 10 years of VTES haven't had at least one extended period (6-12 months) away from the game? >> I'll use myself as an example here. Three years ago, I >> could easily hold my own with any player from Atlanta, Columbia, >> Raleigh-Durham, or Boston (the groups I most commonly played against). I >> was winning drafts in Atlanta, and making finals in constructed events >> consistently, against some of the best competition on the East Coast. In >> the intervening time, a relocation and two kids has _severely_ cut down >> my >> play time, and I know that I'm no where near as polished a player as I >> once >> was... A rating system without an expiration date would show me as a far >> greater threat in the next tournament down the road than is at all >> warranted. > > OK, but you're talking about rust as being the specific reason. No, I'm talking about rust (lack of general game-play experience), _and_ my lack of in-game familiarity with new cards, their text, and their function. And I put more emphasis on the lack of knowledge and experience with the new cards on my 'lower-than-would-be-predicted-from-prior-lifetime-performance' at the New England Qualifier. I can shake off the 'rust' in a few casual games. I cannot generate experience with the new cards, their dynamic and function, and recall their in-game effects within a similar timeframe... > Even rust > flakes off pretty quickly, though - allowing a player to get back to being > about the kind of player he used to be. Are you understanding that one of > the changes you'd have to make is to tune the ELO coefficients DOWN to the > point that a few bad games or even a few bad tournaments won't move your > score all that much? Sure, (assuming you can find anyone to administer your ELO system). > How long would you expect to spend getting back > to your old level? I believe your abilities and experience, in the long > run, is going to be more important than rust. > Thats my point. A person _without experience with newer cards_ is going to play at a level less than is predicted by their rating (given a system that doesn't somehow age their results). > But if you slow down and play on a much reduced level over an indefinite > period of time, It doesn't need to be indefinite. It doesn't even need to be a year... > sure - your skill will truly go down because you don't > play enough. It happens. Just like peoples' skill goes up with more > play. I'd still rather have the possibility for out-of-date ratings than > the significant and certain bias based on participation we now have by > a long shot. > Then go and build the better ratings-trap. After all of this thread, will you be suprised when everyone shrugs, and says the old system was perfectly sevicible? >> You have yet to show (other than the sole example of the Tatu Factor) the >> extent to which the current system is 'horribly biased'. At least Bakija >> is >> willing to crunch a few numbers... > > Bakija's numbers don't mean anything, except that he misunderstands how to > think about the bias. I don't see how the fact that most of the people > on the top 50 list happen to have good VP/game ratios (would was anyone > expecting, anyway) shows non-correlation between number of tournaments > and rating points. If you have a high VP/game ratio or not, you will > still > have a higher score if you attend 12 tournaments than if you attend 8, > a higher still score if you attend 16, and a higher still score if you > attend 20 - at least on average. The bias is that two equal skilled > players, whatever their VP/game ratio, will not be equally rated on > average if one has attended signficantly more tournaments than the > other. THAT'S what the "Tatu factor" is. I have no idea what Peter > thinks it is. And that is an obvious thing, requiring no statistics > to back it up. > >> _You're_ the one with the issue with the >> current system (and have plenty of time to post about it, again and again >> and again), but you can't take the time to dig up the stats to back your >> assertions? > > The only thing I could use statistics for is demonstrate what the > magnitude of the bias is. Well, thats something, isn't it? > And this would depend on various factors > such as how many points do players score on average depending on > how good they are and what size tournaments they attend and mainly, > how much do the finishes vary from one another - which shows how much > of an advantage attending additional tournaments over 8 would be. > > You can't do this solely by digging up statistics kept by VEKN. > You'd also have to do a statistical analysis on the randomness of > finishes varying by skill level (how would one define that?) and > plug that into formulas you'd have to invent around using the top 8 > out of X finishes. I have a Bachelor's degree in mathematics and > I don't have a clue where to start. (Granting I didn't take many > statistics courses.) Do you? Have a Bachelor's in Mathematics? No, merely Chemistry... But then, I'm not the one quintuple-posting about how awful the current system is, am I? > > Fred >

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF3403CD.2179E%pd...@lightlink.com... > Frederick Scott wrote: [ quoted text not captured ] What's kind of funny to me about the analysis you did is that it's more or less recreating the *last* rating system we had before this one - the one where the stats ranked were "VPs per game" and "GWs per game". In that system, the people at the top of the list were typically those who had managed to play at least 10 games while maintaining some extraordinarily high averages in those stats - like winning 80% of their games or whatnot. Anyway - I think my point is that the VPs per game and GWs per game stats may not be all that meaningful either. In fact the most important VPs and GWs (by far) in the current ranking system are the ones in tournament finals - all the rest are relatively less important, except insofar as they get you into those final games. One other thing I noticed is that the GWs per game you calculated are actually pretty substantially variable among players, rather more so than the VPs per game. Since VPs per game were mostly above 1.0, the difference between 1.5 and 1.9 is, what, like a 27% difference. But the difference between 0.30 GWs per game (Jay Kristoff at #8 worldwide) and 0.47 GWs per game (Erik Torstensson at #12 worldwide) is, like, a 57% difference. >> The most obvious flaw (though not the only one) is the same one that >> I've >> pointed at all along: that good players with high VP/Games comparable >> to >> or better than these may be a ways down the list - where you can't >> see >> them just by calculating the Top 27. > > Sure. Because they are winning smaller events. Or not winning that > many > games. Players who do well at "prestige" events (the big ones) are > going to > be ranked higher than someone with similar VP/Game and GW/Game stats. > The > rating is not made by you VP/Game or GW/Game, but by the quality of > the VPs > and Games you win, as abstracted, reasonably, by virtue of the size > and > quality of the event--a VP won in an 8 person local tournament is > worth > less, in terms of ranking, than a VP won in a 60 person Qualifier. > Which is > fine. Well, in most of the rounds, a VP in an 8-person tournament is worth exactly the same, in rating points, as one in a 60-person qualifier. The difference is in the tournament finals, where the tournament coefficient (including qualifier or championship bonus) is applied to the 90/50/30/20/10-point bonuses. Anyway. To me, the current system makes sense inasmuch as it's a "star system" kind of ranking; it greatly rewards winning tournaments compared to anything else, and as such its leader board is basically a list of people who have won a substantial number of good-size tournaments in the last 18 months. I find it an interesting list because, to me, winning tournaments is a pretty good way to rank VTES players without throwing a huge amount of computational infrastructure at the question. It doesn't do so well at ordering rankings for people who aren't big participants in the tournament scene, obviously, and that's something Fred would like it to do better. But I don't think it can do a really good job of that without, again, being more computationally intensive than is practical for VEKN to support. And its slipshod ability to rate everyone is probably good enough for a lot of people, who may like it as a bragging-rights kind of thing but don't take it *too* seriously. Josh you can't rate anyone cause you're unrateable

Joshua Duffin wrote: > What's kind of funny to me about the analysis you did is that it's more > or less recreating the *last* rating system we had before this one - the > one where the stats ranked were "VPs per game" and "GWs per game". Only in the sense that, ya know, it has those stats. I'm not using those stats to indicate any sort of ranking, though, which is what I think Fred seems to think I'm trying to do. The numbers I came up with (at least in my mind) are irrelevant in and of themselves--they are only useful in relation to each other. Sure, like, Bean Peal at #2 has a 0.37 GW/Game where David Fraile at #25 has a 0.39 GW/Game stat. If we were looking only at the GW/Game stat as an indicator, then David Fraile is ranked higher than Ben Peal. But that wasn't why I was coming up with those numbers. It was to look at what everyone had to do to get a good rating, and everyone did more or less the same thing--everyone played around the average number of games (sure, some high, some low, but very few wildly high or wildly low) and scored about the same number of VP's per game (somewhere between 1 and 2) and scored around the same number of game wins (somewhere between 1/5 and 1/2)--no one was, like, winning an average of 3 VP per game, or winning like 4/5ths of the games they played. And only one player in everyone I looked at in 70 had an "artificially" high rating based on a large number of games making up for below average VP and GW performance (ya know, Tatu), indicating (at least in my possibly flawed logic) that the system we have tends to reward good performance over participation. Sure, you *can* get a high score through participation rather than good scoring, but very few people in the system are actually doing that--the vast, vast majority of the players in the system are getting ratings based on similarly good play. >Anyway - I think my point is that the VPs per game > and GWs per game stats may not be all that meaningful either. They clearly aren't, in and of themselves (people with high ratings regularly have lower GW/Game or VP/Game than people with much lower ratings) as the statistics I came up with don't take into account the "quality" of each VP and Game win, nor do they take into account finalist bonuses (i.e. in the numbers I came up with, 2GWs count as only 2GW. In the rating system, 2GW usually also come with a "I get into the finals" bonus). So again, those numbers aren't particularly meaningful. Except to see what kind of performance everyone in the system needs to get into the system, and to see that you generally get high ratings by playing well in an average number of games as opposed to by playing below averagely over a well above average number of games. > One other thing I noticed is that the GWs per game you calculated are > actually pretty substantially variable among players, rather more so > than the VPs per game. Since VPs per game were mostly above 1.0, the > difference between 1.5 and 1.9 is, what, like a 27% difference. But the > difference between 0.30 GWs per game (Jay Kristoff at #8 worldwide) and > 0.47 GWs per game (Erik Torstensson at #12 worldwide) is, like, a 57% > difference. Yep. But they are all somewhere between winning 1 game in 5 and winning 1 game in 2 (about 13 people in the top 50 are in the 40% or better zone, most people are in the 30-40% zone, or about 1 game in 3). There is a variable, sure, but I don't think it is *that* huge, and it is certainly affected by total games played--the more total games someone plays, the more their GW/Game tend to average out--the people with a lot of games tend to have more average GW/Game (Ben Peal with 197 games is at 0.37, where the average is 0.35; Miiko Raimi with 26 games is at 0.61). > Well, in most of the rounds, a VP in an 8-person tournament is worth > exactly the same, in rating points, as one in a 60-person qualifier. > The difference is in the tournament finals, where the tournament > coefficient (including qualifier or championship bonus) is applied to > the 90/50/30/20/10-point bonuses. Correct, but those points are where all the ratings jumps come from--if you get, say, 2GW, 10VP, and 1st in an 8 person tournament, you get, what, about 95 rating points, but if you get that same score in a 60 person tournament, you get something like 350 rating points. > Anyway. To me, the current system makes sense inasmuch as it's a "star > system" kind of ranking; it greatly rewards winning tournaments compared > to anything else, and as such its leader board is basically a list of > people who have won a substantial number of good-size tournaments in the > last 18 months. I find it an interesting list because, to me, winning > tournaments is a pretty good way to rank VTES players without throwing a > huge amount of computational infrastructure at the question. Agreed. > It doesn't do so well at ordering rankings for people who aren't big > participants in the tournament scene, obviously, and that's something > Fred would like it to do better. But I don't think it can do a really > good job of that without, again, being more computationally intensive > than is practical for VEKN to support. And its slipshod ability to rate > everyone is probably good enough for a lot of people, who may like it as > a bragging-rights kind of thing but don't take it *too* seriously. Yep. [ quoted text not captured ]

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF3403CD.2179E%pd...@lightlink.com... > Frederick Scott wrote: >>I agree, a higher than 1-per-game ration is obviously >> good, but how high should it be? What does a difference between two scores >> mean? Is a difference of 0.1 big or small? Who knows, without calculating >> every player's rating, from top to bottom, and then doing statistical >> analysis on it? > > All I'm looking at here, really, is averages in performance--I'm not trying > to come up with "skill" or whatever. The numbers show that most players tend > to cluster around the same average performance level--the average of the top > 50 players in the world is to get about 1.5 VP per game, winning about 1/3rd > of the games they play, over about 115 games. And most players in the top 50 > fit this model--the incidence of people having high ratings from playing > lots of games and not doing so well is insignificant. If by "fit the model", you mean they don't vary so much in terms of VP-per- games as David Tatu does, sure. But that doesn't mean anything in terms of the "Tatu factor". The Tatu factor is just a question of how playing many tournaments affects your current rating. And that works just fine for players of all VP-per-games ratios, except zero of course. I'm totally confused why you've pursued this analysis. > Overall, the vast > majority of players get their rankings by doing the same things that the > other, similarly ranked players do. They play about the same numbers of > games, they win about the same numbers of VPs per game, and they win about > the same percentage of games. There is certainly *some* variation, but > certainly among the top 50 players in the world, they are all doing pretty > much the same thing to get that high ranking. And so...? > >> The most obvious flaw (though not the only one) is the same one that I've >> pointed at all along: that good players with high VP/Games comparable to >> or better than these may be a ways down the list - where you can't see >> them just by calculating the Top 27. > > Sure. Because they are winning smaller events. Or FEWER events?!? Because - although they're winning events at the same ratio as others because they're just as good as others - they've PLAYED fewer events: the Tatu factor. > Again, playing > lots of games to make up for lackluster performance certainly can make a > difference in a single person's rating (although apparently, it doesn't > actually that much). But in terms of players who have a hig score based on > many games disguising low scores overall, there is a total of 1 in the top > 50 (Mr. Tatu). Everyone else in the top 50 got to the top 50 by doing the > same thing--playing about the same number of games, winning about the same > number of VPs per game, and winning about the same percentage of games. The > difference at that point comes down to, more than anything, the "quality" of > the points won--were the VPs and Game Wins from big events with high levels > of competition or were they VPs and GWs from small, local events. Playing > more big events pushes you up where playing small local events keeps you > lower. That quality of the points won may be one thing. Prolificacy is the effect I'm complaining about, however. >> The reason I brought David Tatu is that he doesn't >> seem to have played any better than Charnley by tournament placement yet >> he was rated much higher than Charnley. > > Yes. 'Cause David is the outlier. The only one. The "only one" because that's what you looked at. I used him because he was the most obvious one. So you waltzed off and found a way in which he's sort of unique on the list but that's got nothing to do why Peter Charnley is way down at 27 on top 50 list. You've somehow gotten totally obsessed with a statistic that means nothing to the issue. > *Everyone* has a good chance of raising their ratings by attending more > tournaments. That is part of the system. My point, when I originally posted, is that this is the part of the system which chronically underrates certain players who do not partake of this feature. Keep in mind, this is all relative so try not to get too obsessed with David as I say this. Peter is disadvantaged with respect to _everyone_ who played more and larger tournaments than he did in the course of 18 months. Some people ahead of him have scored well with surprisingly few tournaments. (Nice record, Peter! Btw.) But if you look through list, some people have pretty long lists. Anyone who does can be suspected of being where they are based on the Tatu factor. > The Tatu Factor is meaningless in the sense that it does not have a > significant impact on ratings for some and not others--everyone benefits > from playing more games, and yet everyone (in the top 50) is playing about > the same number of games. Huh? It sure doesn't look like that to me. Compare, for instance, Matt Flint (53 constructed games) with David Wilson (25 constructed games). Does David deserve to be ranked lower than Matt? I think somehow, in your obsession with calculating VP ratios, you've missed the real point of the Tatu factor. Fred

"lehrbuch" <lehr...@gmail.com> wrote in message news:430e8d6d$1...@news.maxnet.co.nz... > Frederick Scott wrote: >> So all of this varies quite a bit and we'd have to do a lot of number >> crunching to understand how much a ninth tournament adds to the >> average player's score... > > The R^2 value is a measure of the correlation between two variables a > value of 1 = strong correlation, a value of 0 = no correlation. > > Using the figures elsewhere of Peter's, if you calculate the R^2 value > of the correlation between Ranking and Games Played it has a value of > 0.09, for the top 50. That is, there is not a strong correlation between > Games Played and Ranking, which would be expected if the Tatu factor was > very significant. > > Additionally, if you calculate the R^2 for the correlation between > Ranking and VP/Game, it has a value of 0.10. That is, Ranking does not > strongly correlate to this measure of "ability", either. > > However, if you calculate the R^2 for the correlation between Ranking > and Wins/Game, it has a value of 0.20. Again this is not a strong > correlation, but nonetheless the Ranking represents those players who > won, more than it represents those players who played lots. > > Therefore, a ranking correlates more to winning games than playing lots. > This seems all good. This is all interesting. But it's hard to know how to take it without some kind of description that better sheds light on how the R^2 value is calculated or what it means. From what you're saying, it does sound like skill in game wins is at least more important than participation, which I agree is good. But it's not very clear how much more important. And whether there might be some other factors that explain the low correlation factors (I'm taking your word for it that they're "low") for all three things. For instance, does the emphasis on making it to and winning the final game screw up the correlation calculations? How about other aspects of the system, like "attendance points", the "top 8" rule, and the size of the tournament? I'm not sure what "figures elsewhere of Peter's" means but if you're using only the leader list from the United States, I'd also speculate that using all players' statistic would show a much stronger correlation between number of games and rating points. In short, the current system may be much more accurate for really, really good players than for mundane players. (Not that it's terribly accurate for really good players.) Fred

On 21 Aug 2005 23:03:21 -0700, <jeff...@pacbell.net> wrote: > BAN ARIKA! ;) ...and PTO. :D -- Bye, Daneel

On Mon, 22 Aug 2005 20:07:17 -0400, Derek Ray <lor...@yahoo.com> wrote: > Oh, quit your fucking sour-grapes whining, Fred. The ranking system > works, whether you like it or not, and whether you're willing to admit > it or not. > > Suck on it, deal with it, and shut the FUCK up about it, OK? > YOU GET TO BE WRONG THIS TIME. Your caps lock is stuck. Time to burn your keyboard. And don't bother gettin a new one, it ain't worth it. ;) -- Bye, Daneel

On Thu, 25 Aug 2005 13:20:21 GMT, David Zopf <david...@snetx.net> wrote: > You seem to be neglecting an aspect of VTES that to me seems both > obvious and critical. Unlike chess, bridge, or any of the other > rated games used as a comparitor in this thread, VTES (both its > rules and its play components) changes over time. It could just as > easily be said that the rating system slides the time period (very > slowly) because the game of VTES today is significantly different > from the game of VTES played 18 months ago. Perhaps a reason you, > in fact, _must_ use participation and a sliding time frame in > rating VTES players is bacause, if you want to rate player skill > in a VTES game today or next week, then the results from some > longer time period ago aren't that valuable a predictor of future > performance. Okay, good point, but I basically disagree. The key word is adaptation. That's simply a component to being a skilled VTES player. Just like strategic thinking, tactical talent and people manipulation. Now, it could be argued that as the rules change, the game is basically becoming something completely different - like shifting from Game A (original Jyhad) to Game B (a future, hypothetic game that will be played when VTES is finally discontinued, hopefully not too soon). But I think that those two games are not that different in principle. -- Bye, Daneel

On Tue, 23 Aug 2005 23:26:27 +0200, Stefan Ferenci <nos...@thankyou.com> wrote: > Frederick Scott wrote: > >> >> I wasn't claiming they were. I was just pointing out that 115th is a >> pretty >> low ranking for a future Continental Champion - *especially* for one who >> clearly didn't "come out of nowhere" but was a previous Continental >> Championship >> finalist. >> >> Worthless. >> >> Fred > > rankings are not supposed to predict the future, they are supposed to > evaluate the past (18 month to be precise) *these* rankings. Not rankings in general. Sure, this system does something. If you don't mind circular arguments, you can just say that this system is supposed to do exactly what it does, so since it does what it is supposed to, it is working fine. The only downside to this way of thinking is that it can be applied to each and every system concieavable. -- Bye, Daneel

In message <opsv5o32...@news.chello.hu>, Daneel <dan...@eposta.hu> writes: >*these* rankings. Not rankings in general. Sure, this system does > something. If you don't mind circular arguments, you can just say > that this system is supposed to do exactly what it does, so since > it does what it is supposed to, it is working fine. The only > downside to this way of thinking is that it can be applied to each > and every system concieavable. Not really. You're confusing intent with implementation. It is perfectly possible for a system's designer to intend for it to do X, Y and Z but to screw up painfully. -- James Coupe PGP Key: 0x5D623D5D YOU ARE IN ERROR. EBD690ECD7A1FB457CA2 NO-ONE IS SCREAMING. 13D7E668C3695D623D5D THANK YOU FOR YOUR COOPERATION.

On Sat, 27 Aug 2005 09:32:26 +0100, James Coupe <ja...@zephyr.org.uk> wrote: > In message <opsv5o32...@news.chello.hu>, Daneel <dan...@eposta.hu> > writes: >> *these* rankings. Not rankings in general. Sure, this system does >> something. If you don't mind circular arguments, you can just say >> that this system is supposed to do exactly what it does, so since >> it does what it is supposed to, it is working fine. The only >> downside to this way of thinking is that it can be applied to each >> and every system concieavable. > > Not really. You're confusing intent with implementation. No, not really. Also check the part you snipped (what I replied to). > It is perfectly possible for a system's designer to intend for it to do > X, Y and Z but to screw up painfully. Yes it is. But you confuse cause with explanation... ;) -- Bye, Daneel

In message <opsv5xii...@news.chello.hu>, Daneel <dan...@eposta.hu> writes: >On Sat, 27 Aug 2005 09:32:26 +0100, James Coupe <ja...@zephyr.org.uk> >wrote: > >> In message <opsv5o32...@news.chello.hu>, Daneel <dan...@eposta.hu> >> writes: >>> *these* rankings. Not rankings in general. Sure, this system does >>> something. If you don't mind circular arguments, you can just say >>> that this system is supposed to do exactly what it does, so since >>> it does what it is supposed to, it is working fine. The only >>> downside to this way of thinking is that it can be applied to each >>> and every system concieavable. >> >> Not really. You're confusing intent with implementation. > >No, not really. Also check the part you snipped (what I replied to). Why? I was responding to your claim that it could be applied to each and every system conceivable, not anything else you wrote. [ quoted text not captured ]

David Zopf wrote: > I _did_ specifically point out that I was tlaking about those with a lack of > current-time experience (say someone slacks off playing for a year, or > doesn't buy in to a particular set, etc.), so I'll go along with the fact > that there must be an additional reason besides change itself. If a player > keeps buying and playing through the change, he'll be as well off as he ever > was. But honestly, how many among us in the past 10 years of VTES haven't > had at least one extended period (6-12 months) away from the game? You know Dave, all my experience in chemistry is with bleeding edge 1950's technology. But I'm pretty sure even these newfangled masspectrahoozits give ya some dead time. So you *could* wander on over to #VTES on IRC and just, you know, shoot the shit. There are enough sharp people hanging out there (mostly) talking about cards that it will keep you up to speed with where the game's at. And you get to say hey to people, like me, xian, ben, matt hirsch, etc et al! It's been helpful to me because Durham is going through kind of a V:TES slump at the moment; but mostly because of #vtes and deckbot, I feel as up to speed with the last set as I think I would have when we were cranking out one or two nights of V:TES a week. And, you know.... it's IRC: "Nothing to see here, move along."

Frederick Scott wrote: > If by "fit the model", you mean they don't vary so much in terms of VP-per- > games as David Tatu does, sure. But that doesn't mean anything in terms > of the "Tatu factor". The Tatu factor is just a question of how playing > many tournaments affects your current rating. And that works just fine > for players of all VP-per-games ratios, except zero of course. I'm > totally confused why you've pursued this analysis. To indicate that the validity of the ratings, across the board, are not compromised by some people playing tons of games while other people are playing a few games. Yes. You get a higer rating by playing more games. That is the design of the system. And to get a high rating, you have to play lots of games. And consequently, everyone who does have a high rating has, more or less, played lots of games. And 49 of the top 50 players have all hit the top 50 by virtue of doing, across lots of games, pretty much uniformly well. Indicating, again, at least in my possibly flawed logic, that the system, in reality, rewards solid play over attendance. Attendance is certainly a factor, but in a practical sense, playing well is more important than showing up, assuming that everyone shows up about the same amount, and they do. > The "only one" because that's what you looked at. I used him because he > was the most obvious one. So you waltzed off and found a way in which > he's sort of unique on the list but that's got nothing to do why Peter > Charnley is way down at 27 on top 50 list. You've somehow gotten totally > obsessed with a statistic that means nothing to the issue. Charnley was down on the list 'cause he wasn't doing that well. He'll be much higher on the list now that he has done better. Remember--the system does not measure skill. It measures performance. Previous to winning the NAC, Charnley may have been a skillful player, but he wasn't performing all that well, and as the system does not measure skill, but instead measures performance. Now this his performance is better, his rating will be better. > My point, when I originally posted, is that this is the part of the system > which chronically underrates certain players who do not partake of this > feature. That is 'cause the system rewards you for showing up to more tournaments. This is not a flaw. It is a feature. Luckily, though, everyone who is doing well in the system is doing well by virtue of about the same amount of effort. > Keep in mind, this is all relative so try not to get too obsessed > with David as I say this. Peter is disadvantaged with respect to > _everyone_ who played more and larger tournaments than he did in the > course of 18 months. Everyone is disadvantaged with respect to people who play more and larger tournaments. This is not a flaw. It is a feature. Of a system that measures performance and not skill. > Huh? It sure doesn't look like that to me. Compare, for instance, Matt > Flint (53 constructed games) with David Wilson (25 constructed games). > Does David deserve to be ranked lower than Matt? I think somehow, in > your obsession with calculating VP ratios, you've missed the real point > of the Tatu factor. Umm, wha? I think you have your numbers wrong. Those 2 look like: 82. Matt Flint 132 games; 36 GW; 177.5 VP; 0.27 GW/Game; 1.34 VP/Game 97. David Wilson 83 games; 21 GW; 106.5 VP; 0.25 GW/Game; 1.28 VP/Game Matt has played more games that David, yes. Their performance per game is virtually identical. Why is Matt ranked higher than David? 'Cause Matt has probably played bigger (and by extrapolation, harder to win at) events. And possibly Matt has been in the finals more than David (I'll go check that later) and possibly won more events than David. 'Cause that is how the system works. They both have enough games to cover the necessary 8 tournaments. They both have enough games to have twice or more those 8 games. They both are playing just as well. Matt is higher 'cause the system is designed to rank some folks higher due to what events they play in and what events they win, as bigger events are worth more. [ quoted text not captured ]

I wrote: > And > possibly Matt has been in the finals more than David (I'll go check that > later) and possibly won more events than David. 'Cause that is how the > system works. I looked at their records, and saw: Matt Flint's best 8 games: 1st in 9 player 2nd in 9 player 2nd in 18 player 2nd in 17 player 3rd in 15 player 4th in 24 player 4th in 24 player Qualifier 5th in 25 player David Wilson's best 8 games: 1st in 12 player 1st in 10 player Qualifier 2nd in 19 player 3rd in 8 player 4th in 12 player 17th in 41 player 46th in 48 player David hasn't played 8 events in the last 18 months, so he is behind an event. Perhaps if he had a good score in one more event, he'd be higher than Matt, as David has 2 first place finishes, where Matt only has 1. But as they system works, Matt has a full 8 events to fall back on, that are good. David does not. Matt gets a higher rating. But still, Matt's higer rating is not so much higher that it is, like, completely egrigious, based on his large spread of events--they are, what, 82 and 97? [ quoted text not captured ]

On Sat, 27 Aug 2005 12:05:07 +0100, James Coupe <ja...@zephyr.org.uk> wrote: > In message <opsv5xii...@news.chello.hu>, Daneel <dan...@eposta.hu> > writes: >> On Sat, 27 Aug 2005 09:32:26 +0100, James Coupe <ja...@zephyr.org.uk> >> wrote: >> >>> In message <opsv5o32...@news.chello.hu>, Daneel <dan...@eposta.hu> >>> writes: >>>> *these* rankings. Not rankings in general. Sure, this system does >>>> something. If you don't mind circular arguments, you can just say >>>> that this system is supposed to do exactly what it does, so since >>>> it does what it is supposed to, it is working fine. The only >>>> downside to this way of thinking is that it can be applied to each >>>> and every system concieavable. >>> >>> Not really. You're confusing intent with implementation. >> >> No, not really. Also check the part you snipped (what I replied to). > > Why? I was responding to your claim that it could be applied to each > and every system conceivable, not anything else you wrote. *sigh* So for what systems concievable, in your highly esteemed opinion, can the following application of (circular) logic not be used: The system does something. I assume that the system is supposed to do exactly what it does. Based on that assumption I conclude that the system does exactly what it is supposed to. -- Bye, Daneel

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Peter D Bakija wrote: > Frederick Scott wrote: > >>If by "fit the model", you mean they don't vary so much in terms of VP-per- >>games as David Tatu does, sure. But that doesn't mean anything in terms >>of the "Tatu factor". The Tatu factor is just a question of how playing >>many tournaments affects your current rating. And that works just fine >>for players of all VP-per-games ratios, except zero of course. I'm >>totally confused why you've pursued this analysis. > > To indicate that the validity of the ratings, across the board, are not > compromised by some people playing tons of games while other people are > playing a few games. > > Yes. You get a higer rating by playing more games. That is the design of the > system. And to get a high rating, you have to play lots of games. And Neither of these statements are technically true, and it's this which is, as always, the core of the confusion. To get a high rating, you must play no more than 8 tournaments. If you perform well in all of them, you will have a high rating; guaranteed. Also, playing more tournaments does NOT, despite all feeble protestations to the contrary, guarantee a higher rating. You still have to perform better than you have in previous tournaments, or your rating simply won't go up once you've passed the minimum of 8. To use an analogy, I can get as many at-bats as I want against Randy Johnson, and my batting average is not mysteriously going to rise to .400 after a certain point -- because I can't hit at the major-league level, and simply increasing my number of at-bats ain't gonna change it. If I play 12 tournaments, get fives in two, tens in two more, and score ~50 rating points in all the rest, I'm going to have around 400 rating points total. If I now play a 13th tournament and get 50 rating points, my rating is not going to significantly change. Nor will it change if I get fived again. In fact, my rating can only change if I somehow manage to perform BETTER than I have in the past. To put it in plain English: I play some more tournaments, I do better than I have in the past, and my rating goes up. How good is that? - -- Derek insert clever quotation here -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFDEHkYtQZlu3o7QpERAhU5AKCXenFEvcE997OUPywhWLrQ18MmXgCfYXXP +EJc2otmc/kq3AJ5+NrWJdE= =VVeA -----END PGP SIGNATURE-----

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 [ quoted text not captured ] Your obsession is showing again. - -- Derek insert clever quotation here -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFDEHqAtQZlu3o7QpERAqVOAKC+74OwXehFdba/8f0nNUxcg9jIIACeI765 UTFq4AODofHezFRt+vO6INA= =lBZ4 -----END PGP SIGNATURE-----

Derek Ray wrote: > Neither of these statements are technically true, and it's this which > is, as always, the core of the confusion. True. And there seems to be lots of confusion :-) > To get a high rating, you must play no more than 8 tournaments. If you > perform well in all of them, you will have a high rating; guaranteed. Correct. > Also, playing more tournaments does NOT, despite all feeble > protestations to the contrary, guarantee a higher rating. You still > have to perform better than you have in previous tournaments, or your > rating simply won't go up once you've passed the minimum of 8. Also correct. > To use an analogy, I can get as many at-bats as I want against Randy > Johnson, and my batting average is not mysteriously going to rise to > .400 after a certain point -- because I can't hit at the major-league > level, and simply increasing my number of at-bats ain't gonna change it. Man. If I had any understanding of baseball at all, this would probably make sense. But as soon as someone starts making a baseball analogy, all I can hear is "Blah, blah, blah, Ginger, blah, blah." Which is a problem I have. Not 'cause baseball is bad, but 'cause I have some sort of complete mental block against it or something. Although a story I heard on NPR once about the guy who wrote the book analyzing how the A's did so well for so liitle money was kind of fascinating. But anyway... > If I play 12 tournaments, get fives in two, tens in two more, and score > ~50 rating points in all the rest, I'm going to have around 400 rating > points total. If I now play a 13th tournament and get 50 rating points, > my rating is not going to significantly change. Nor will it change if I > get fived again. In fact, my rating can only change if I somehow manage > to perform BETTER than I have in the past. Correct again. > To put it in plain English: I play some more tournaments, I do better > than I have in the past, and my rating goes up. How good is that? Fantastic. Which is why I think the system is a reasonable one. [ quoted text not captured ]

Daneel wrote: > Your caps lock is stuck. Time to burn your keyboard. And don't bother > gettin a new one, it ain't worth it. ;) Not that, like, I'm trying to make trouble or anything, but you are responding to a post, like, a week old. And not even in a constructive way. And from Derek, who, like, you are always busting into. I'm just saying. [ quoted text not captured ]

In message <opsv57ks...@news.chello.hu>, Daneel <dan...@eposta.hu> writes: >So for what systems concievable, in your highly esteemed opinion, can > the following application of (circular) logic not be used: > >The system does something. I assume that the system is supposed to do > exactly what it does. Based on that assumption I conclude that the > system does exactly what it is supposed to. You're working from a flawed principle - that people are doing something ludicrously stupid. People are extrapolating certain principles and testing them against what is both sensible and useful. Note that some of the people involved with such things have expressed their views about, for example, simplicity and cost of running an ELO system! From this, we can establish things without just looking at the numbers. However, if you would bother with, well, anything other than stamping, posturing and point-scoring against Derek Ray, you'd understand the difference between strategy and tactics, and how these can be examined from an established system and its results. For example, in one of my other interests (politics), I have an interest in voting systems. It is extremely possible to examine a voting system for its intent but also to find its flaws in real life situations. As a specific example of that, I have spent time examining the Single Transferable Vote system. The general system shows certain points, the implementation when it hits real life shows other points (which may counter the original points of the system, as flaws turn up because behaviour isn't as expected[0] or other variables weren't taken into account for some reason), and also what it prioritises in certain interesting border cases[1]. Examining all sorts of interesting votes - in particular, the Republic of Ireland[2] - can show up all the interesting ways that the theory breaks down in real life. One specific example here is that the system obviously intends to reward people for doing well regularly. However, people (such as Peter) are examining the real life implications for The Tatu Factor - to see if that actually holds in real life, because in real life it could work in interestingly different ways. That *you* don't understand the difference between circular logic and extracting principles and implications is not a reason for the rest of us to not do so. If it troubles you, feel free not to post. [0] Something like 8% of votes in Australia go to the top name on the ballot, irrespective of party, from an interesting combination of STV and compulsory voting. [1] For example, whilst STV (as maintained by the Electoral Reform Society) attempts to give everyone an equal say in as far as that can be achieved, it fails to do so if you weren't in the latest incoming batch of votes. That is, if you vote early on in your list for a candidate and later on a surplus is transferred from that candidate, you won't transfer whereas someone who was only just transferred to that candidate (from a surplus or exclusion) will be. This is an interesting trade-off between practicality (transferring the fractional surplus which becomes extremely small in some cases e.g. a few vote surplus from an early candidate, which ensures that each transferred vote is worth only a tiny fraction, which is then transferred in a later surplus transfer), against the wish to give people a say and prevent tactical voting. [2] They have it down to a fine art. Parties promoting a specific platform of candidates (typically, all their own party, of course) can have this down to telling people exactly how to vote and when - it varies at different times of day - in order to get the right candidates elected, and in the right order, and so on. [ quoted text not captured ]

On Sat, 27 Aug 2005 11:26:52 -0400, Peter D Bakija <pd...@lightlink.com> wrote: > Daneel wrote: > >> Your caps lock is stuck. Time to burn your keyboard. And don't bother >> gettin a new one, it ain't worth it. ;) > > Not that, like, I'm trying to make trouble or anything, but you are > responding to a post, like, a week old. And not even in a constructive > way. The point being what, exactly? I. Working a lot -> less time to read -> weekend more time -> read lots of stuff -> post some stuff. II. Non-constructive post by notorious poster -> poke fun (in possibly non-constructive way). > And from Derek, who, like, you are always busting into. I'm just saying. Dunno, I don't really care who writes what. If someone is terribly stupid, I often feel like pointing that out (even if it won't really change a thing; I kind of feel like I have the right to respond if someone thought they had the right to post something stupid, and I went through the trouble to read it in hopes of finding content). I can't really help it if certain people are terribly stupid more often than others. I mean, looking at the facts, of course. I can imagine that people who usually post inane things have an inane theory of their own for being occasionally busted into. Really, it goes for everyone: post as you would posted to. -- Bye, Daneel

Daneel wrote: > The point being what, exactly? > > I. Working a lot -> less time to read -> weekend more time -> read lots > of stuff -> post some stuff. > > II. Non-constructive post by notorious poster -> poke fun (in possibly > non-constructive way). Sure, sure. But when you poke fun at a post that is a week old, even with real life getting in the way, it looks like you are just making trouble unecessarily, rather than poking fun. Like, if you posted the same thing the day the original post was made, no one would have noticed. But whatever Derek posted vanished from our minds, and then a week later, here you are cracking wise on it. I'm not saying you should be cracking wise--just choose your timing better. Yeah, it might have taken you a week to get to the post to crack wise, but it might be better, in the long run, to save the cracking of wise for discussions that are current, rather than a week old. I don't think anyone would have a problem with a cogent post about something a week old, but the cracking wise over a discussion that is a week old just seems, I dunno, kind of lame. [ quoted text not captured ]

On Sat, 27 Aug 2005 16:54:38 +0100, James Coupe <ja...@zephyr.org.uk> wrote: > In message <opsv57ks...@news.chello.hu>, Daneel <dan...@eposta.hu> > writes: >> So for what systems concievable, in your highly esteemed opinion, can >> the following application of (circular) logic not be used: >> >> The system does something. I assume that the system is supposed to do >> exactly what it does. Based on that assumption I conclude that the >> system does exactly what it is supposed to. > > You're working from a flawed principle - that people are doing something > ludicrously stupid. No. I pointed out what I considered to be flawed reasoning. The example I presented was intended as a negative example outlining the specific pitfalls of reasoning inherent to the point I refuted. You know, the point you snipped when you chimed in. ;) Yes, the reasoning I gave was purposefully extreme. So then you go into the trouble of responding, and point out that my (purposefully extreme) example is flawed, because it isn't absolute. I try to drive you back to seeing the (purposefully extreme) example in context and understanding it thusly... Which you fail to do, claiming that my (purposefully extreme) example stands on its own and is wrong. So I then challenge you to prove that my (purposefully extreme) example is wrong, and you respond with, "Oh, but your example is extreme, so I can't prove its wrong; therefor [bla-bla] and [more bla-bla]". I mean, this was probably quite pointless. Unless, of course, you just felt like picking on someone, hairsplitting a little and hiding a little name-calling there as well. Well, newsflash, I don't really give a rat's ass about namecalling. Forums in general are filled with so much trash that I kind of became immune to any frustration the prevalent inanity and malignancy characteristic to the darker side of human nature, crystallized by the opportunity to employ a medium that mostly or completely avoids any personal accountability, could cause to the unsuspecting reader. If you write something memorable with regard to content, I'll take you seriously and agree or debate*. Otherwise, I'll just think the post was written in an inane or malignant way and deserves no real attention; depending on the tone I'll probably ignore it, or at times point out the inanity or malignancy in a way I presume will be understood by whomever the post was written by. * Even disregarding anything you might've written earlier - I try to see posts, not posters. If I do note a poster, I try to do it for something positive. -- Bye, Daneel

On Sat, 27 Aug 2005 12:24:47 -0400, Peter D Bakija <pd...@lightlink.com> wrote: > Sure, sure. But when you poke fun at a post that is a week old, even with > real life getting in the way, it looks like you are just making trouble > unecessarily, rather than poking fun. Like, if you posted the same thing > the > day the original post was made, no one would have noticed. But whatever > Derek posted vanished from our minds, and then a week later, here you are > cracking wise on it. > > I'm not saying you should be cracking wise--just choose your timing > better. > Yeah, it might have taken you a week to get to the post to crack wise, > but > it might be better, in the long run, to save the cracking of wise for > discussions that are current, rather than a week old. I don't think > anyone > would have a problem with a cogent post about something a week old, but > the > cracking wise over a discussion that is a week old just seems, I dunno, > kind > of lame. Well, on the one hand I don't quite understand what you mean. I sometimes have piles of usenet posts waiting on my computer, and when I have time, I read whole threads. It may be, for that very reason, that I do not really follow how old something is - except when I want to cross-link a post or something to the local forums, and I need to go through google. I kind of see though where you are coming from. From an "oh, that's so last friday" point of view, sort of. So, opinion noted, data processing scheduled. ;) -- Bye, Daneel

Daneel wrote: > I kind of see though where you are coming from. From an "oh, that's so > last friday" point of view, sort of. So, opinion noted, data processing > scheduled. ;) Yeah, that is completely what I mean. Like, sure, not everyone reads the NG every day (or every 5 minutes, for that matter :-), but even then, at a certain point, it is reasonable to assume that discussion posts that are a week old, are, ya know "totally last week", or whatever, and usually, they are long forgotten. Sometimes, if someone sees an old post and has something interesting to say about it, it might restart an interesting conversation, which is good. On the other hand, if someone sees an old post and wants to crack wise on it a week later, that tends to come off as just making trouble. But then, the same could be said about, ya know, handing out helpful advice :-) [ quoted text not captured ]

Derek Ray wrote: > To use an analogy, I can get as many at-bats as I want against Randy > Johnson, and my batting average is not mysteriously going to rise to > .400 after a certain point -- because I can't hit at the major-league > level, and simply increasing my number of at-bats ain't gonna change it. If your chance of hitting a pitch from Randy Johnson is not zero, and your batting average is based on a finite subset that includes all your best hits rather than the whole set of the times you bat (as analogous to the eight-in-eighteen sample), then as the number of times you bat against him approaches infinity, your batting average will approach 1.000. The more tournaments a player plays in, the higher his or her eight-in-eighteen rating is likely to be. I don't see this as a problem with the existing system. In my opinion, the system recognizes players' prominence in the tournament scene, something that is based partly on skill and partly on luck and partly on attendance. Tatu should be rated highly because he's a fixture of the scene.

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Peter D Bakija wrote: > Derek Ray wrote: > >>To use an analogy, I can get as many at-bats as I want against Randy >>Johnson, and my batting average is not mysteriously going to rise to >>.400 after a certain point -- because I can't hit at the major-league >>level, and simply increasing my number of at-bats ain't gonna change it. > > Man. If I had any understanding of baseball at all, this would probably make > sense. But as soon as someone starts making a baseball analogy, all I can > hear is "Blah, blah, blah, Ginger, blah, blah." Which is a problem I have. To put what's necessary into perspective: Batting average of .400 means that you hit safely 4 out of 10 times. .400 is outstanding, world-class, best-ever. .300 is considered "damn fine hittin'". - -- Derek insert clever quotation here -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFDEhGntQZlu3o7QpERAi+iAJwIBBZZ64kq+U5F7mokORjyzwBnvACg83DG pSwAhtGS16NzEXZAhKr2xdM= =n94c -----END PGP SIGNATURE-----

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Emmit Svenson wrote: > Derek Ray wrote: > >>To use an analogy, I can get as many at-bats as I want against Randy >>Johnson, and my batting average is not mysteriously going to rise to >>.400 after a certain point -- because I can't hit at the major-league >>level, and simply increasing my number of at-bats ain't gonna change it. > > If your chance of hitting a pitch from Randy Johnson is not zero, and > your batting average is based on a finite subset that includes all your > best hits rather than the whole set of the times you bat (as analogous > to the eight-in-eighteen sample), then as the number of times you bat > against him approaches infinity, your batting average will approach > 1.000. A better analogy would be using "batting average per game", then, and give bonus points to batters who finish in the top 5. However, it doesn't need to be a perfect analogy to illustrate the point quite adequately -- which is that until you get better, your rating will NOT go up significantly simply by attending tournaments until the cows come home. There's just no way around it; you're capped at your eight best. > The more tournaments a player plays in, the higher his or her > eight-in-eighteen rating is likely to be. But it will not be a significant change unless that player begins performing much better than they have in the past. > I don't see this as a problem with the existing system. In my opinion, > the system recognizes players' prominence in the tournament scene, > something that is based partly on skill and partly on luck and partly > on attendance. Tatu should be rated highly because he's a fixture of > the scene. If we're going to leave all rating systems out of it and go with personal impressions -- Tatu should be rated highly because he really IS that good when he bothers to play decks that aren't completely goofy. He spends a lot of time playing decks that are, frankly, goofy; hence the zero-VP performances. But that's one thing the rating system was deliberately designed to do -- NOT penalize skilled players for experimenting with new strategies. - -- Derek insert clever quotation here -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFDEhLktQZlu3o7QpERApBqAKDfKkY9gihZvUJdaCv3bCWGgBKZCACgx15+ OC575YHIcrF2zBa4iJvyXQY= =JUxB -----END PGP SIGNATURE-----

Derek Ray wrote: > To put what's necessary into perspective: > > Batting average of .400 means that you hit safely 4 out of 10 times. > > .400 is outstanding, world-class, best-ever. > .300 is considered "damn fine hittin'". Excellent. Thanks. So then, like, your point was that as you are not a superstar baseball player, hitting a million times isn't going to give you a high at bat average. You'll still be below average. Check! [ quoted text not captured ]

Frederick Scott wrote: > "lehrbuch" <lehr...@gmail.com> wrote in message [snip] >> Therefore, a ranking correlates more to winning games than playing lots. >> This seems all good. > > This is all interesting. But it's hard to know how to take it without > some kind of description that better sheds light on how the R^2 value > is calculated or what it means. http://en.wikipedia.org/wiki/Correlation_coefficient > From what you're saying, it does sound like skill in game wins is at > least more important than participation, which I agree is good. But > it's not very clear how much more important. And whether there might > be some other factors that explain the low correlation factors (I'm > taking your word for it that they're "low") for all three things. For comparison the correlation coefficient between US firearms sales and the US murder rate is 0.35 (1985-1993). Compared to this the correlation is low. Note, of course, that correlation does not equal causation. For another comparison the correlation between two sets of 50 random numbers generated by my computer (PC, using Excel) varies between about 0.1 and 0.00001. Which probably says something about MS's "random number" generator. > For instance, does the emphasis on making it to and winning the final > game screw up the correlation calculations? How about other aspects > of the system, like "attendance points", the "top 8" rule, and the > size of the tournament? The fact that the statistics are for a *career* while ranking is for the *last 18 months* is, I would guess, the primary reason why the correlations are low. > I'm not sure what "figures elsewhere of Peter's" means but if you're > using only the leader list from the United States... The figures that Peter had seem to be the top 50 from the white-wolf page these appear to be the world leaders rather than the US. If you are interested you can do the analysis on other data for yourself. The function will be built in to any typical computer statistics package. -- * lehrbuch (lehr...@gmail.com)

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Peter D Bakija wrote: > Derek Ray wrote: > >>To put what's necessary into perspective: >> >>Batting average of .400 means that you hit safely 4 out of 10 times. >> >>.400 is outstanding, world-class, best-ever. >>.300 is considered "damn fine hittin'". > > Excellent. Thanks. So then, like, your point was that as you are not a > superstar baseball player, hitting a million times isn't going to give you a > high at bat average. You'll still be below average. Check! Yes. It occurs to me that I've left out two additional facts: 1) Randy Johnson is a star pitcher for the Yankees and has been known to throw fastballs at 95+ MPH routinely. 2) I played golf from the time I could hold a club until I was 14, and as such the politest description that can be given to my baseball swing is "worthless." (Neither of these facts do much more than uphold the above statement, to wit, that the odds of me hitting safely 4 out of any given 10 times against Randy Johnson are directly proportional to the odds of me finding a way to bribe Randy Johnson to throw underhand during the process.) - -- Derek insert clever quotation here -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFDEnmEtQZlu3o7QpERAq/WAJ9rxYkkYT+gGHklHKHGd5bgL8X96gCcDcdV lNFmSHEyjwQnlQqktutziY4= =PyAO -----END PGP SIGNATURE-----

"Gregory Stuart Pettigrew" <ethe...@sidehack.gweep.net> wrote > > Wes nearly won Shadow Twin Constructed on Friday. That was certainly an odd game. A lesson in deal-making and deal-keeping, methinks. Cheers, WES

"David Cherryholmes" <david.che...@gmail.com> wrote in message news:11h0mg4...@corp.supernews.com... > David Zopf wrote: > >> I _did_ specifically point out that I was tlaking about those with a lack >> of current-time experience (say someone slacks off playing for a year, or >> doesn't buy in to a particular set, etc.), so I'll go along with the fact >> that there must be an additional reason besides change itself. If a >> player keeps buying and playing through the change, he'll be as well off >> as he ever was. But honestly, how many among us in the past 10 years of >> VTES haven't had at least one extended period (6-12 months) away from the >> game? > > You know Dave, all my experience in chemistry is with bleeding edge 1950's > technology. But I'm pretty sure even these newfangled masspectrahoozits > give ya some dead time. So you *could* wander on over to #VTES on IRC and > just, you know, shoot the shit. Thanks, but unfortunately IRC is banned at my workplace, due to abuses by another workmate (oddly, usenet is still considered OK...) I should probably try to get on #VTES at night (once the kiddos are asleep). I get the odd evening game in on VTES Online Beta, but it seems my available time doesn't often coincide with others on that format, either. *shrug* I think I'll be all set when VTES Online gets released to its full audience... 'Sides, as you know, my responsiblities now fall 1/2 in chemistry, and 1/2 in sales/marketing, so I don't get that free day-time while running reactions as much as I once did. Even still, thanks for thinking of me and my plight. > There are enough sharp people hanging out there (mostly) talking about > cards that it will keep you up to speed with where the game's at. Heh. Unfortunately, I'm a pretty solid empiricist, and I only seem to learn well by doing (rather than discussing), paired with a healthy dose of failing ;-). Really, my point was only about how people who wander away from the game (for whatever the reason), and then come back, even over a relatively short time period, will be mis-represented by a system which doesn't age their results in some way. And I think that if you're looking to the "middle of the pack" as far as a rated group goes, that you're more likely to need to account for this, in order to rank folks accurately. DaveZ AW

> For comparison the correlation coefficient between US firearms sales and > the US murder rate is 0.35 (1985-1993). Compared to this the > correlation is low. Note, of course, that correlation does not equal > causation. What I like is the use of selective, outdated information to back up an arguement. Especially since firearm sales has grown signifigantly since concealed weapons laws passed in many states, and the murder rate in the U.S. has dropped like a rock since 1995 (roughly 50%). Back to your regularly scheduled programming.... Comments Welcome, Norman S. Brown, Jr XZealot Archon of the Swamp

XZealot wrote: [lehrbuch] >> For comparison the correlation coefficient between US firearms sales and >> the US murder rate is 0.35 (1985-1993). Compared to this the >> correlation is low. Note, of course, that correlation does not equal >> causation. > > What I like is the use of selective, outdated information to back up an > arguement. Absolutely. > Especially since firearm sales has grown signifigantly since concealed > weapons laws passed in many states ... [snip] That's nice. The data I used was the first sensible looking data that I could find with a quick google search, that was semi-meaningful. The point was to demonstrate that a correlation coefficient of 0.1 or 0.2 is low, compared to the sorts of correlations that people normally argue about --- 0.35 is also low. > Back to your regularly scheduled programming.... Indeed. -- * lehrbuch (lehr...@gmail.com)

"James Coupe" <ja...@zephyr.org.uk> wrote in message news:0Y$egIRaU...@gratiano.zephyr.org.uk... > In message <opsv5o32...@news.chello.hu>, Daneel <dan...@eposta.hu> > writes: >>*these* rankings. Not rankings in general. Sure, this system does >> something. If you don't mind circular arguments, you can just say >> that this system is supposed to do exactly what it does, so since >> it does what it is supposed to, it is working fine. The only >> downside to this way of thinking is that it can be applied to each >> and every system concieavable. > > Not really. You're confusing intent with implementation. > > It is perfectly possible for a system's designer to intend for it to do > X, Y and Z but to screw up painfully. Assuming there was some profound conceptual intent for what it should do that is distinct from just doing whatever the hell it turns out to do. Fred

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF35E73B.217F4%pd...@lightlink.com... > Frederick Scott wrote: > >> If by "fit the model", you mean they don't vary so much in terms of VP-per- >> games as David Tatu does, sure. But that doesn't mean anything in terms >> of the "Tatu factor". The Tatu factor is just a question of how playing >> many tournaments affects your current rating. And that works just fine >> for players of all VP-per-games ratios, except zero of course. I'm >> totally confused why you've pursued this analysis. > > To indicate that the validity of the ratings, across the board, are not > compromised by some people playing tons of games while other people are > playing a few games. Well, you haven't shown that. Two players can have exactly the same VP- per-game ratio and one can have a much higher rating because he's played more than the other one. > Charnley was down on the list 'cause he wasn't doing that well. Charnly was down on the list due to lack of events. If he'd done better in the events he was in, he'd also be higher. But it's abundantly clear that by just entering more tournaments and getting a spread of results from them that are comparable to the ones he already has would BY ITSELF move him up the list a ways. >> Huh? It sure doesn't look like that to me. Compare, for instance, Matt >> Flint (53 constructed games) with David Wilson (25 constructed games). >> Does David deserve to be ranked lower than Matt? I think somehow, in >> your obsession with calculating VP ratios, you've missed the real point >> of the Tatu factor. > > Umm, wha? I think you have your numbers wrong. Those 2 look like: > > 82. Matt Flint 132 games; 36 GW; 177.5 VP; 0.27 GW/Game; 1.34 VP/Game > 97. David Wilson 83 games; 21 GW; 106.5 VP; 0.25 GW/Game; 1.28 VP/Game Um, you're looking at *career* games. I was looking at the number of games their rating is based from. For that, you have to call up their individual records for the last 18 months and add them. > Matt has played more games that David, yes. Their performance per game is > virtually identical. Why is Matt ranked higher than David? 'Cause Matt has > probably played bigger (and by extrapolation, harder to win at) events. He has played some larger tournaments than David has, yes. I don't think any of those tournaments are in his top 8, though. His top points come from a 9 and a 10 player tournament. Some other likely sources are tournaments in the 20-25 player range. David's 40-50 player tournaments weren't good showing for them either but they're on his score 'cause he only has 7 tournaments. In short, it's both. And I don't care which it is, neither thing is skill. Both players have some skill and show it in some-but-not-all of their tournaments. If David had played as many and as large of tournaments as Matt, it seems likely he'd overtake Matt, the difference only being around 50 points. > Matt is higher 'cause the system > is designed to rank some folks higher due to what events they play in and > what events they win, as bigger events are worth more. In this case, it seems fairly easy to generalize that Matt is higher solely due to opportunity. Fred

"Derek Ray" <lor...@yahoo.com> wrote in message news:XI2dnQnKkv-...@giganews.com... > To get a high rating, you must play no more than 8 tournaments. If you > perform well in all of them, you will have a high rating; guaranteed. > > Also, playing more tournaments does NOT, despite all feeble > protestations to the contrary, guarantee a higher rating. No one was talking about "guaranteed higher rating" in any given situation. Just that if people who play more tournaments do better under the system. Just because you can postulate that a player's 16 worst tournaments might happen to be his last 16 tournaments out of 24 changes none of this. > To use an analogy, I can get as many at-bats as I want against Randy > Johnson, and my batting average is not mysteriously going to rise to > .400 after a certain point -- because I can't hit at the major-league > level, and simply increasing my number of at-bats ain't gonna change it. That analogy is completely flawed. They don't use the average from your eight best at bat solely to compute your batting average. If they did, players with more at-bats with have higher averages - to a limit of 1.000 of course. And understanding that if you had no chance of hitting a major league pitcher then you won't raise your .000 average no matter how many at-bats you had. > To put it in plain English: I play some more tournaments, I do better > than I have in the past, and my rating goes up. How good is that? 1. It's an oversimplied summation. 2. The flaws are in the generalizations. Fred

> 1. It's an oversimplied summation. 2. The flaws are in the > generalizations. Hey Fred. . . You write lots of e-mails. You talk a lot. You clearly have an opinion not shared by lots of folks. You claim the current rating system isn't great. Take your time spent here and create a better one. Address all the myriad concerns brought up in these threads. Report. If you don't have enough interest to do this after this extended discussion, then clearly you don't agree that it needs to be done. Because I *can* say this. . . it doesn't seem that you have *any* shortage of time. Personally, as one of them ivory tower math folks, I think it is elegantly written and models quite a few things. There might be a minor tweak or two I'd make to it to try to normalize one's performance across some reasonable measure, but I haven't worked out exactly what I'd do or how (or if) it would be better. Ankur Gupta Prince of West Lafayette "Corn always have free-for-alls, so ELO doesn't apply."

Frederick Scott wrote: > Charnly was down on the list due to lack of events. If he'd done better > in the events he was in, he'd also be higher. But it's abundantly clear > that by just entering more tournaments and getting a spread of results from > them that are comparable to the ones he already has would BY ITSELF move > him up the list a ways. He could have been higher on the list in one of two ways: -He could have done better in the events he played in. -He could have gone to other events, and done well at those. Either one works for getting a higher rating. No one is arguing that going to more events doesn't give you a higher rating (I was arguning, in my lengthy figure analysis, that the rankings aren't overtly skewed by people who played huge numbers of games while not doing that well, which they aren't; I was not arguing that playing lots of games doesn't help your ranking--it clearly does). The system is based on that. 'Cause it measures your performance at the events you go to. > Um, you're looking at *career* games. I was looking at the number of games > their rating is based from. For that, you have to call up their individual > records for the last 18 months and add them. Oh, yeah, ok. > In short, it's both. And I don't care which it is, neither thing is skill. Which is appropriate, as the rating system doesn't measure skill. It doesn't try. It measures performance. > Both players have some skill and show it in some-but-not-all of > their tournaments. If David had played as many and as large of tournaments > as Matt, it seems likely he'd overtake Matt, the difference only being > around 50 points. Sure. So he needs to go play those tournaments and do well at them if he wants to overtake Matt. > In this case, it seems fairly easy to generalize that Matt is higher solely > due to opportunity. I don't know if "opportunity" is actually the issue in this case--I suspect that David (for example) had as much opportunity to play in events as Matt (by "opportunity" I mean "enough available events")--but likely moot, as there isn't really a way to investigate how many tournaments David (just using him as an example here--not really looking for a number or anything) had access to. Yeah, Matt has played more events in the same span of time. The system, which measures performance in a number of events over a span of time (and not skill) is designed to give a benefit to people who go to more events. That is part of the design of the system. The more events you go to, the more get measured. [ quoted text not captured ]

Frederick Scott wrote: [snip] > In short, it's both. And I don't care which it is, neither thing is > skill. Both players have some skill and show it in some-but-not-all of > their tournaments. If David had played as many and as large of tournaments > as Matt, it seems likely he'd overtake Matt, the difference only being > around 50 points. Of course, as has been pointed out before, this is because the rating system *rates performance at tournaments*; which is, of course, one possible definition of "skill". It does NOT rate any other definition of "skill". No one thinks that it does rate another definition of "skill". Clearly, unskilled players cannot get "high" ratings because by definition they do not perform well at tournaments --- assuming that whatever definition of "skill" you would like to use is somehow connected to winning games or VPs. Equally, a player that is skilled but does not attend enough(any) tournaments and/or always plays silly bugger decks at tournaments will also not get a high rating. There is nothing surprising about this. The system rates tournament performance because: a) that's what it's meant to do; b) tournament performance *is* measurable, unlike many other definitions of skill; c) a rating based on past tournament performance is a handy (but obviously not infallible) guide to future tournament performance. There is no argument here. You do not appear to be saying anything useful. -- * lehrbuch (lehr...@gmail.com)

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Frederick Scott wrote: (nothing important) I wasn't talking to you, Fred. I don't consider you either intelligent enough or open-minded enough to understand what was said. All I see from you is a person hellbent on intentionally misunderstanding the rating system's intent, purpose, and functionality for reasons known only to himself. As such, I'm not about to waste my valuable time trying to argue with you; others seem to have enough patience to do so, and I'll leave it to them. As far as I'm concerned, you can still get fucked. - -- Derek insert clever quotation here -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFDFR6DtQZlu3o7QpERAiWQAKDzlHFKNDs6atUMhwPiz++E+36SNgCeN7jl vX74JdCGyayN64C0CEa+b98= =/qur -----END PGP SIGNATURE-----

On Tue, 30 Aug 2005, Peter D Bakija wrote: > Sure. So he needs to go play those tournaments and do well at them if he > wants to overtake Matt. You know Peter. . . . I'm hoping this conversation goes on long enough to eventually include every single vekn-registered player in the world. So far I count about 60 players having been used as examples. How much time do you give it? A week? Two? Ankur Gupta Prince of West Lafayette "Apparently not a 'factor' in the ratings like Tatu. :)"

[ quoted text not captured ] I think I can see where Fred is coming from a bit more clearly than apparently most of the people who keep on arguing with him. Or maybe not. Nevertheless, the way I see it, Fred does the dirty work of reminding folks about the facts. You typical "VTES rating" discussion goes like this: Forumite A: The rating system is good because [...] and [...] and [...]. Fred: Well, possibly, but let's not confuse things, and keep in mind that it does not directly measure skill. Forumite B: Yeah, well, fuck you, you're an asshole, and the system does measure skill, because the sky is blue and the grass is green. Fred: ... you are wrong. [explain why the current ranking system does not measure skill, and how it takes attendance into consideration]. Forumite C: Hmm... you're right... Then make a new system. Fred: Well, I kind of just want to point out the facts. I don't necessarily mind the way things are as long as we know exactly what is what and we keep calling a spade a spade. Forumite D: Wow, long thread, what I really like in the current rating system is how it measures skill. Fred: *sight* Forumite B: Yeah, and Ferd isnt' untellugent unough to anderstand dis! #@%˘!"%@!! ...so, I kind of sympathize with Fred. ;) -- Bye, Daneel

Daneel wrote: > You typical "VTES rating" > discussion goes like this: (snipped entertaining meta-lampoon of usenet) Or, ya know, he has a lengthy discussion with people who don't do that. And just point out that the system was never designed to measure skill, which Fred seems to think it was. [ quoted text not captured ]

Ankur Gupta wrote: > You know Peter. . . . I'm hoping this conversation goes on long enough to > eventually include every single vekn-registered player in the world. So > far I count about 60 players having been used as examples. How much time > do you give it? A week? Two? No, no, I think we are up to, um, 70? -Top 50 worldwide -Top 10 NE US -Bottom 10 of 50 NE US -Then Matt and David I think two people have been represented in two groups (Matt and Lance Shoppe?) See, both Fred and I have, what we like to call, "the steely tenacity", so it is certainly possible we'll get everyone:-) [ quoted text not captured ]

On Wed, 31 Aug 2005 08:54:24 -0400, Peter D Bakija <pd...@lightlink.com> scrawled: >Ankur Gupta wrote: > >> You know Peter. . . . I'm hoping this conversation goes on long enough to >> eventually include every single vekn-registered player in the world. So >> far I count about 60 players having been used as examples. How much time >> do you give it? A week? Two? > >No, no, I think we are up to, um, 70? >-Top 50 worldwide >-Top 10 NE US >-Bottom 10 of 50 NE US >-Then Matt and David > >I think two people have been represented in two groups (Matt and Lance >Shoppe?) > >See, both Fred and I have, what we like to call, "the steely tenacity", so >it is certainly possible we'll get everyone:-) OOh! Do me! Do me! salem http://www.users.tpg.com.au/adsltqna/VtES/index.htm (replace "hotmail" with "yahoo" to email)

"Ankur Gupta" <agu...@cs.duke.edu> wrote in message news:Pine.GSO.4.62.05...@eenie.cs.duke.edu... > Take your time spent here and create a better one. Address all the myriad > concerns brought up in these threads. Report. If you don't have enough > interest to do this after this extended discussion, then clearly you don't > agree that it needs to be done. The premise doesn't justify the conclusion. (There are lots of reasons one might not have interest in doing it that don't imply disagreement that it needs to be done.) But never mind. You know the situation as well as I do. You're a prince. You're aware that Steve Wieck implemented the current system based on discussion from the conclave. However I feel about that, there's no point in actively campaigning for a change. However, I am free to point out problems with the current system on this forum and to dispute incorrect statements made about it as I go. Which is what I am doing. Fred

"Peter D Bakija" <pd...@lightlink.com> wrote in message news:BF3B200B.21930%pd...@lightlink.com... > Daneel wrote: > >> You typical "VTES rating" >> discussion goes like this: > > (snipped entertaining meta-lampoon of usenet) > > Or, ya know, he has a lengthy discussion with people who don't do that. And > just point out that the system was never designed to measure skill, which > Fred seems to think it was. Yea. I've had pretty reasonable discussion with you and and a few others at times. Daneel does make one good point: when you're one guy arguing with multiple others making varied and profuse different points with which you disagree there is a tendency to make a lot of posts, sound like you're repeating yourself (in fact, the "other side" taken as a unit repeats itself continually - it just doesn't look like it so much because the statements get repeated by different mouths who say the same thing in different ways), and get caught in a crossfire by new posters who suddenly jump in when they misunderstand the point you were making to someone else because they don't have the full context. I don't mind. It's the lot of someone who wants to defend an unpopular point of view. It does get kind of stupid when certain people critize you for stuff like repeating yourself or posting a lot, though. Fred

"lehrbuch" <lehr...@gmail.com> wrote in message news:4315...@news.maxnet.co.nz... > The system rates tournament performance because: a) that's what it's > meant to do; b) tournament performance *is* measurable, unlike many > other definitions of skill; c) a rating based on past tournament > performance is a handy (but obviously not infallible) guide to future > tournament performance. > > There is no argument here. You do not appear to be saying anything useful. I am saying that rating "performance" - as opposed to rating skill - has no coherent meaning. A while back I made a post in response to something Derek Ray posted which proposed three different systems which could all be said to measure "performance" as much as the existing system. I also conjectured three different mythical players with their mythical records. Under each different "performance rating" system, the three players rated completely differently, each the king of one of the systems. What purpose could rating people under a paradigm like this possibly serve? The word Peter is using, "performance", has no useful meaning. "Skill" in this context at least has meaning, even if there's no perfect way to measure it. Fred

In message <2yLRe.156320$E95.26687@fed1read01>, Frederick Scott <nos...@no.spam.dot.com> writes: >Under each different "performance rating" system, the three players rated >completely differently, each the king of one of the systems. What purpose >could rating people under a paradigm like this possibly serve? the same is true under many "skill based" systems. The coefficients of an ELO system will differ, the probability curves can be drawn differently, not all skill based systems are ELO-based anyay, the rises and falls will differ, different players will come out well, some systems may treat draws in a different way to other systems. What purpose could rating people under a paradigm like this possibly serve? That different systems produce differing results applies to many, many systems, whether they rank performance or "skill". -- James Coupe PGP Key: 0x5D623D5D YOU ARE IN ERROR. EBD690ECD7A1FB457CA2 NO-ONE IS SCREAMING. 13D7E668C3695D623D5D THANK YOU FOR YOUR COOPERATION.

"James Coupe" <ja...@zephyr.org.uk> wrote in message news:oy$5qMTfp...@gratiano.zephyr.org.uk... > In message <2yLRe.156320$E95.26687@fed1read01>, Frederick Scott > <nos...@no.spam.dot.com> writes: >>Under each different "performance rating" system, the three players rated >>completely differently, each the king of one of the systems. What purpose >>could rating people under a paradigm like this possibly serve? > > the same is true under many "skill based" systems. The coefficients of > an ELO system will differ, the probability curves can be drawn > differently, not all skill based systems are ELO-based anyay, the rises > and falls will differ, different players will come out well, some > systems may treat draws in a different way to other systems. > > What purpose could rating people under a paradigm like this possibly > serve? > > That different systems produce differing results applies to many, many > systems, whether they rank performance or "skill". The comparison between the two situations is inapt. Ultimately, one understands the meaning of "skill", even if the various ELO systems may measure it well or poorly in a given context for various reasons. You can have a debate about how accurate the system is, if it isn't accurate (in the speaker's opinion) why isn't it accurate, and what could be done to make the system more accurate. By comparison, one has no idea whatsoever of what is meant by "performance" in this context except by consulting the system itself. Therefore, what "performance" only means is whatever the arbritrary formula that adjudicates between participation and results happens to spit out. You can NOT debate whether the system is "accurate" because there is no intuitive notion of what you're trying to get from it with which you can compare the actual results. To charge that the system is "inaccurate", as I was suggesting in the case of Peter Charnley for instance, leaves me vulnerable to Peter's counterargument that it effectively does just exactly what it was meant to do. Which is nothing, AFAICS. Fred

> "Ankur Gupta" <agu...@cs.duke.edu> wrote in message > news:Pine.GSO.4.62.05...@eenie.cs.duke.edu... >> Take your time spent here and create a better one. Address all the myriad >> concerns brought up in these threads. Report. If you don't have enough >> interest to do this after this extended discussion, then clearly you don't >> agree that it needs to be done. > > The premise doesn't justify the conclusion. (There are lots of reasons one > might not have interest in doing it that don't imply disagreement that it > needs to be done.) But never mind. You know the situation as well as I I disagree. What sorts of reasons might those be? If any of them have to do with not having enough time, I think it's patently true that that's false. If your reasons have to do with not understanding the math, then you're clearly out of your league in the discussion to begin with. If they have to do with not having access to the data, well, that's solveable. The key is. . . do you have the motivation to show that you're right, or are you going to talk until the cows go home? I'm not saying one way or another whether you're right. Maybe you are. But show us that a system similar to what was in place before can still be good. Right now. . . no one is convinced. And we're kinda giving this one a try. And you know, it seems to be doing alright and stuff. > do. You're a prince. You're aware that Steve Wieck implemented the > current system based on discussion from the conclave. However I feel > about that, there's no point in actively campaigning for a change. Isn't there? And if there isn't, why open your mouth about it to begin with? If you're defeatist enough to think that nothing you say could possibly change anyone's mind. . . . Why argue/discuss? > However, I am free to point out problems with the current system on this > forum and to dispute incorrect statements made about it as I go. Which > is what I am doing. You're of course free to do so. It's just stupid if you feel that there's no possible consequence that could stem from it. And if you do feel there's a way for something to come of it, you need some proof. If you want something done. . . [ quoted text not captured ]

"Ankur Gupta" <agu...@cs.duke.edu> wrote in message news:Pine.GSO.4.62.05...@eenie.cs.duke.edu... >> "Ankur Gupta" <agu...@cs.duke.edu> wrote in message >> news:Pine.GSO.4.62.05...@eenie.cs.duke.edu... >>> Take your time spent here and create a better one. Address all the myriad >>> concerns brought up in these threads. Report. If you don't have enough >>> interest to do this after this extended discussion, then clearly you don't >>> agree that it needs to be done. >> >> The premise doesn't justify the conclusion. (There are lots of reasons one >> might not have interest in doing it that don't imply disagreement that it >> needs to be done.) But never mind. You know the situation as well as I > > I disagree. What sorts of reasons might those be? If any of them have to do with not having enough time, I think it's patently > true that that's false. Cool assertion. I wish there were more things in this world that were patently true that they were false. :-} Anyway, it's one thing to take snippets of time out of your day to argue with people when you love arguing. It's quite another thing to, for instance, spend a bunch of time developing software to resolve a pretty formidible problem that implementing a different system would take. The software takes far more time and requires a lot more resources than just a newsposter and "the steely tenacity". (I suspect it's more boredom than tenacity but what the hell. The latter sounds better...)\ > If your reasons have to do with not understanding the math, then you're clearly out of your league in the discussion to begin > with. No, I understand the math just fine. The issue with ELO has to do with how to recalculate ELO results when an error is discovered in previous results or when old, hitherto unentered, results suddenly come in. > The key is. . . do you have the motivation to show that you're right, or are you going to talk until the cows go home? I think you're WAYyyyyyy underestimating how much work such a thing would be. It's a different order of magnitude than just making some posts. >> do. You're a prince. You're aware that Steve Wieck implemented the current system based on discussion from the conclave. >> However I feel about that, there's no point in actively campaigning for a change. > > Isn't there? And if there isn't, why open your mouth about it to begin with? Why does anyone open their mouth here? To change peoples' opinions. However, in case you didn't catch the implication, I feel that's a different thing than actively campaigning for a change in systems. The one might be a necessary precursor to the other, the current state of things being what they are. But they're still different animals. >> However, I am free to point out problems with the current system on this forum and to dispute incorrect statements made about it >> as I go. Which is what I am doing. > > You're of course free to do so. It's just stupid if you feel that there's no possible consequence that could stem from it. And if > you do feel there's a way for something to come of it, you need some proof. > > If you want something done. . . I think there are a number of things in the above paragraph which are just your opinions. Obviously, I disagree with them. Fred

> Anyway, it's one thing to take snippets of time out of your day to argue > with people when you love arguing. It's quite another thing to, for > instance, spend a bunch of time developing software to resolve a pretty > formidible problem that implementing a different system would take. > The software takes far more time and requires a lot more resources than > just a newsposter and "the steely tenacity". (I suspect it's more > boredom than tenacity but what the hell. The latter sounds better...)\ I'm a computer scientist. I can't imagine it could take so long. Let's define vague, misleading terms in the above: bunch of time (quantified below for your convenience), formidable (impossible? difficult in which way?), far more time (how much more?), and resources (what are they?). My opinion, based on you know, being a computer scientist: Obfuscation of the fact that you just don't wanna do it. >> If your reasons have to do with not understanding the math, then you're >> clearly out of your league in the discussion to begin with. > > No, I understand the math just fine. The issue with ELO has to do with > how to recalculate ELO results when an error is discovered in previous > results or when old, hitherto unentered, results suddenly come in. How about just don't worry about old results coming in? That's an issue with deployment *once it has already been decided* that the system represents what we want to represent. Here's what I'm saying: take the static data that's already present. Freeze it. Compute ratings. Compare. Relatively easy. You're trying to argue that the rating system of ELO is in some way superior to the one we're currently using. Establish THAT first, and then let's see how to deal with the above problems. Let's see that people are ranked according to some skill rather than some nebulous garbage which is our current rating system. I'm curious to see whether it works. >> The key is. . . do you have the motivation to show that you're right, >> or are you going to talk until the cows go home? > > I think you're WAYyyyyyy underestimating how much work such a thing > would be. It's a different order of magnitude than just making some > posts. Yeah? You spend, what, an hour a day here? Typing 100 wpm with thinking time, reading time, and time to concoct arguments should put you right at that. The above project couldn't possibly take more than say. . . 20 hours of work. Heck, I could safely assign this as a class project to my students. So. . . you could come back in a working month and give us the needful. (No, my students are not available as monkeys to do this work.) >>> do. You're a prince. You're aware that Steve Wieck implemented the >>> current system based on discussion from the conclave. However I feel >>> about that, there's no point in actively campaigning for a change. >> >> Isn't there? And if there isn't, why open your mouth about it to begin >> with? > > Why does anyone open their mouth here? To change peoples' opinions. > However, in case you didn't catch the implication, I feel that's a > different thing than actively campaigning for a change in systems. The > one might be a necessary precursor to the other, the current state of > things being what they are. But they're still different animals. So. . . what you're saying is that your arguments are largely without any sort of goal at all other than the ego-stroking satisfaction of changing an irrelevant person's opinion. Man, I wish life worked like this. Sign me up. >>> However, I am free to point out problems with the current system on >>> this forum and to dispute incorrect statements made about it as I go. >>> Which is what I am doing. >> >> You're of course free to do so. It's just stupid if you feel that >> there's no possible consequence that could stem from it. And if you do >> feel there's a way for something to come of it, you need some proof. >> >> If you want something done. . . > > I think there are a number of things in the above paragraph which are > just your opinions. Obviously, I disagree with them. Noted. [ quoted text not captured ]

Frederick Scott wrote: > It does get kind of stupid when > certain people critize you for stuff like repeating yourself or posting > a lot, though. Well, that is always stupid, regardless of the context. It's like these people have never seen the internet before or something :-) [ quoted text not captured ]

In message <cmMRe.156336$E95.12500@fed1read01>, Frederick Scott <nos...@no.spam.dot.com> writes: >By comparison, one has no idea whatsoever of what is meant by >"performance" in this context except by consulting the system itself. >Therefore, what "performance" only means is whatever the arbritrary formula >that adjudicates between participation and results happens to spit out. Erm, it seems somewhat, erm, biased to claim that a performance based rating is some "arbitrary formula" but that a skill-based rating isn't. Because you've just had to pull whatever you decide is "skill" out of your ass when designing such a formula, it is just another "arbitrary formula". Note also that a skill based ELO system very often doesn't rank skill appropriately. In a relatively closed group, if I regularly beat you but am not that much better than you, my score will eventually climb to whatever level the formula decrees it can. (Since your score would fall, I would eventually reach a level where I can't improve against you.) How would my reaching that asymptote even begin to compare my skill to someone actively playing against much better players who hovers around the same level? >You can NOT debate whether the system is "accurate" because there is >no intuitive notion of what you're trying to get from it with which you >can compare the actual results. To charge that the system is "inaccurate", >as I was suggesting in the case of Peter Charnley for instance, leaves me >vulnerable to Peter's counterargument that it effectively does just exactly >what it was meant to do. Which is nothing, AFAICS. If you really think the system was designed to do "nothing", you're completely ignoring the system. You may not like what it does. You may not want it to do what it does. But it very, very clearly is set out to do certain things. Very simple basic points can be derived from its simplicity, its expiring time-frame and its favouring large tournaments and players who reach the final table. These have been discussed at length. You don't like what they do? That's just fine. That's still not "nothing". Sure, you don't like this. And: WE KNOW IT RANKS PERFORMANCE NOT SKILL. YOU DON'T NEED TO KEEP TELLING US. But to say it does "nothing" just opens you to charges that Derek Ray has levelled - that you're close- minded and intent on just banging your drum rather than actually engaging critically in a discussion. [ quoted text not captured ]

"salem" <salem_ch...@hotmail.com> wrote in message news:5nueh1toivl4011g6...@4ax.com... > OOh! Do me! Do me! > Heh. You should put that in your sig file; " PDB did me... with his steely tenacity." DaveZ Atom Weaver

David Zopf wrote: > Heh. You should put that in your sig file; " PDB did me... with his steely > tenacity." Oh, I'm a do-er, alright... [ quoted text not captured ]

"Ankur Gupta" <agu...@cs.duke.edu> wrote in message news:Pine.GSO.4.62.05...@eenie.cs.duke.edu... >> Why does anyone open their mouth here? To change peoples' opinions. However, in case you didn't catch the implication, I feel >> that's a different thing than actively campaigning for a change in systems. The one might be a necessary precursor to the other, >> the current state of things being what they are. But they're still different animals. > > So. . . what you're saying is that your arguments are largely without any sort of goal at all other than the ego-stroking > satisfaction of changing an irrelevant person's opinion. Nope. How you come to such conclusions is a complete mystery to me. Apparently, you've arrived at some philosophy about argumentation where the truth of a point of view is somehow directly connected to whatever random task another person may choose to propose to the speaker. So, for instance, if I suggest that the homeless aren't taken care of well enough in this country, to determine truth, an observer could (for instance) challenge me to single-handedly build every homeless person in the entire country a structure for them to dwell in. Failing that, I must admit that I am somehow "wrong". What you're talking about in this sub-thread, if it held any water, would be the passion of my concern that the system is wrong and that some kind of suffering is going on which needs to be alleviated. There is no suffering here. I'm just saying the system is stupid. That's all. If people like it being stupid, we're totally fine. If I can't get many people to agree that it's stupid, why should I care about crafting a fix for it? People have what they want; what I would do would be a waste of time. The notion that I need to produce physical proof of how an alternate solution might work is just something you've worked yourself up to in your own mind. I see no need for it. I have all the proof I need and that anyone should need that the current system is a stillborn wretch. Replacing it with something else - if anything - is a completely different and currently pointless debate. Fred

"James Coupe" <ja...@zephyr.org.uk> wrote in message news:6ZQNQTcu...@gratiano.zephyr.org.uk... > In message <cmMRe.156336$E95.12500@fed1read01>, Frederick Scott > <nos...@no.spam.dot.com> writes: >>By comparison, one has no idea whatsoever of what is meant by >>"performance" in this context except by consulting the system itself. >>Therefore, what "performance" only means is whatever the arbritrary formula >>that adjudicates between participation and results happens to spit out. > > Erm, it seems somewhat, erm, biased to claim that a performance based > rating is some "arbitrary formula" but that a skill-based rating isn't. Again, I point to the fact that three different "performance" systems yield three different results. How would state that one is a better result than either of the others. How would you know which is best? "Skill" has a concrete and simple definition. At least, under the theorectical ELO system, it is defined by a player's capacity to score more victory points in a game than a given other player. That may not be everyone's idea of what "skill" is, but at least it's easy and intuitive to describe. > Because you've just had to pull whatever you decide is "skill" out of > your ass when designing such a formula, it is just another "arbitrary > formula". Nope. It is attempt to quantify an understandable, universal abstract concept - unlike "performance". > Note also that a skill based ELO system very often doesn't > rank skill appropriately. In a relatively closed group, if I regularly > beat you but am not that much better than you, my score will eventually > climb to whatever level the formula decrees it can. (Since your score > would fall, I would eventually reach a level where I can't improve > against you.) How would my reaching that asymptote even begin to > compare my skill to someone actively playing against much better players > who hovers around the same level? You state that it's a "relatively" closed group - not a completely closed group. Therefore, some crossplay is occurring, at least on an occasional basis. Therefore, the number of points contained with group should be appropriate to the general population. If the group, as a whole, is not as good as the general population, points will leak out during the small amount of crossplay just fine - and everyone in the group should be rated approximately correctly. Therefore, there will not be "much better players who hover around the same level". > If you really think the system was designed to do "nothing", you're > completely ignoring the system. Then describe what it does in a simple, conceptual manner. And in doing so, please explain the relationship between various "performance" measuring systems (the existing one and the others I describe in my response to Derek): which one is the "accurate" system and what is the problem with the other systems that make them inaccurate? You can't do it. You're describing the "velophotometer" I described in my original post many weeks ago. An attempt to measure two or more quantities and describe the result with a single number yields something compeltely worthless. It's crap. It's meaningless. > WE KNOW IT RANKS PERFORMANCE NOT SKILL. > YOU DON'T NEED TO KEEP TELLING US. But to say it does "nothing" just > opens you to charges that Derek Ray has levelled - that you're close- > minded and intent on just banging your drum rather than actually > engaging critically in a discussion. Fine. Answer the questions above. Fred

>> So. . . what you're saying is that your arguments are largely without >> any sort of goal at all other than the ego-stroking satisfaction of >> changing an irrelevant person's opinion. > > Nope. How you come to such conclusions is a complete mystery to me. > > What you're talking about in this sub-thread, if it held any water, > would be the passion of my concern that the system is wrong and that > some kind of suffering is going on which needs to be alleviated. There > is no suffering here. I'm just saying the system is stupid. That's > all. If people like it being stupid, we're totally fine. If I can't > get many people to agree that it's stupid, why should I care about > crafting a fix for it? People have what they want; what I would do > would be a waste of time. > > The notion that I need to produce physical proof of how an alternate > solution might work is just something you've worked yourself up to in > your own mind. I see no need for it. I have all the proof I need and > that anyone should need that the current system is a stillborn wretch. > Replacing it with something else - if anything - is a completely > different and currently pointless debate. My basic point was: Perhaps you'd be better off showing the "proof . . . that anyone *should* need that the current system is a stillborn wretch" (emphasis mine) if you were to you know, address the concrete system that exists. One way to convince people of its fallacies is with explicit evidence that another method works better. Given the amount of time you've spent on this thread, that seems entirely within reason to ask for. That you're not interested in doing it suggests something to me (as per the argumentation discussion), but you're right that there's no compulsion for you to feel similarly. Ankur

"Ankur Gupta" <agu...@cs.duke.edu> wrote in message news:Pine.GSO.4.62.05...@eenie.cs.duke.edu... >> The notion that I need to produce physical proof of how an alternate >> solution might work is just something you've worked yourself up to in >> your own mind. I see no need for it. I have all the proof I need and >> that anyone should need that the current system is a stillborn wretch. >> Replacing it with something else - if anything - is a completely >> different and currently pointless debate. > > My basic point was: > > Perhaps you'd be better off showing the "proof . . . that anyone *should* > need that the current system is a stillborn wretch" (emphasis mine) if you > were to you know, address the concrete system that exists. One way to > convince people of its fallacies is with explicit evidence that another > method works better. Given the amount of time you've spent on this thread, > that seems entirely within reason to ask for. I have given my proof a number of times. It's not a simple thing and different people attack it in different ways reflecting what particular things with which each doesn't agree. But IMHO, I have defended it just fine. "YMMV" is self-evident. > That you're not interested in doing it suggests something to me (as per > the argumentation discussion), but you're right that there's no compulsion > for you to feel similarly. I think for whatever reason, you've come to the conclusion that this task is a lot easier than I think it is and that probably makes a huge difference in our contrasting ways of looking at that point. (And not just the task itself but it seems likely to me that the operation of the resulting software to the point that it would demonstrate anything worthwhile would take some significant effort.) But aside from that, it just seems to me that the question of whether I'm willing to take it on is pretty much an orthagonal issue to the rightness or wrongness of my PoV. Fred