Optimal seatings changes

Please disregard those results, for some of them are wrong due to a flaw in the validation. From now on, I strongly recommand not to use the version 1.5e but version 1.5f : Excel: www.vekn.net/images/stories/downloads/thearchon1.5f.xlsx LibreOffice: www.vekn.net/images/stories/downloads/thearchon1.5f.ods I've been working recently on the 2R+F optimal seatings of the archons, and I've found some more "optimal" seatings than the current ones. The search has been done through exhaustive computations so I doubt there will be better results. Of course, there's more than one seating that is optimal, but they're all equivalent. For instance, for 10 players: T1: 4 9 2 7 1 T2: 10 3 5 6 8 and T1: 4 8 2 6 1 T2: 10 5 9 3 7 are the same in respect to rules 1 to 8. They both violate rule 2 (2. No pair of players share a table through all two rounds, when possible. (N/A in some 2R event.)). Starting transfers can't be equitably distributed (rule 8) because there are 28 transfers to divide between 10 players. Nevertheless, in the first case, we have an absolute deviation of 0,72: 1, 2, 3, 4, 6, 10 have 5 transfers | 5, 8 have 7 transfers | 7, 9 have 6 transfers In the second case, an absolute deviation of 0,72 too: 1, 2, 4, 6, 8, 10 have 5 transfers | 3, 9 have 7 transfers | 5, 7 have 6 transfers. Basically, some players swap their positions. We'll choose arbitrarily one of the optimal seatings as the official one. The current optimal seating for 10 players is: T1: 10 1 4 7 3 T2: 5 6 9 2 8 Which gives us an absolute deviation of 1,28: 1, 6 have 3 transfers | 2, 7 have 6 transfers | 3, 4, 8, 9 have 7 transfers | 5, 10 have 5 transfers which is worse (we have 2 players with 3 transfers). The optimal seatings will be updated in the next version of the archon, here's the current "best" optimal seatings: (NB: in some cases, rule "3. Available VPs are equitably distributed." is violated too, when the tables have different sizes. The optimal seating is the one with the smaller absolute deviation on the vps). 10 players: 4 9 2 7 1 | 10 3 5 6 8 Rule 2 KO (player 1 shares a table with player 2): 24 (once), 8 (twice) Rule 8 KO. Absolute deviation is: 0.72 => 1, 2, 3, 4, 6, 10 have 5 transfers | 5, 8 have 7 transfers | 7, 9 have 6 transfers (Current seating has an absolute deviation of 1.28) 11 players: no calculated, it should be avoided anyway 12 players: 4 11 9 5 | 8 3 1 10 | 12 7 2 6 (fixed in 1.5f) Rule 2 KO (player 1 shares a table with player 3): 30 (once), 3 (twice) Rule 8 KO. Absolute deviation is: 0.333333333333333 => 1, 4, 6, 7, 8, 9, 10, 11 have 5 transfers | 2, 12 have 6 transfers | 3, 5 have 4 transfers (Current seating has an absolute deviation of 0.5) 13 players: 4 13 6 12 7 | 3 8 2 10 | 11 5 9 1 (fixed in 1.5f) Reverted back to 1.5a seating, until we find a better seating. 14 players: 5 3 12 7 11 | 9 13 2 8 1 | 14 10 4 6 (fixed in 1.5f) Rule 2 KO (player 1 shares a table with player 2): 42 (once), 5 (twice) (Current seating is: 40 once, 6 twice) Rule 3 KO. Absolute deviation is: 0,476190476190476 => 1, 4, 6, 8, 11, 12, 13, 14 have 9 VP | 2, 3, 5, 7, 9, 10 have 10 VP (Current seating has the same absolute deviation) Rule 8 KO. Absolute deviation is: 0,836734693877551 => 1, 3 have 4 transfers | 2, 12, 14 have 6 transfers | 4, 5, 6, 7, 8, 11, 13 have 5 transfers | 9 have 7 transfers | 10 have 8 transfers (Current seating has the same absolute deviation) 15 players: 9 13 7 11 1 | 14 5 12 6 2 | 15 4 10 8 3 (fixed in 1.5f) Rule 2 KO (player 2 shares a table with player 5): 48 (once), 6 (twice) Rule 8 KO. Absolute deviation is: 0,72 => 1, 6, 7, 9, 11, 12, 13, 14, 15 have 5 transfers | 2, 4, 5 have 6 transfers | 3, 8, 10 have 7 transfers (Current seating has an absolute deviation of 1.09) 16 players: 16 11 6 1 | 12 15 2 5 | 8 3 14 9 | 4 7 10 13 All rules are OK! (Current seating is OK too!) 17 players: 5 3 7 14 10 | 4 12 15 6 | 9 16 11 1 | 17 13 2 8 (fixed in 1.5f) Rule 2 KO (player 3 shares a table with player 5): 54 (once), 1 (twice) (Current seating is KO, 46 once, 5 twice) Rule 3 KO. Absolute deviation is: 0,617647058823529 => 1, 2, 4, 7, 10, 14 have 9 VP | 3, 5 have 10 VP | 6, 8, 9, 11, 12, 13, 15, 16, 17 have 8 VP (Current seating has a better absolute deviation: 0,5 => 1, 2, 3, 4, 5, 6, 8, 11, 13, 17 have 9 VP | 7, 9, 10, 12, 14, 15, 16 have 8 VP) Rule 8 KO. Absolute deviation is: 0,311418685121107 => 1, 2, 3, 4, 5, 6, 7, 9, 10, 11, 12, 14, 15, 16, 17 have 5 transfers | 8 have 7 transfers | 13 have 6 transfers (Current seating has an absolute deviation of 0.93) 18 players: 9 5 7 15 11 | 4 13 16 6 1 | 10 17 12 2 | 18 8 14 3 (fixed in 1.5f) Rule 2 KO (player 1 shares a table with player 4): 60 (once), 2 (twice) (Current seating has 48 once, 8 twice) Rule 3 KO. Absolute deviation is: 0,555555555555556 => 1, 4, 5, 6, 7, 9 have 10 VP | 2, 3, 8, 10, 11, 13, 15, 16 have 9 VP | 12, 14, 17, 18 have 8 VP (Current seating has a better absolute deviation of 0,1111) Rule 8 KO. Absolute deviation is: 0,518518518518519 => 1, 4, 6, 7, 8, 9, 10, 11, 12, 13, 15, 16, 17, 18 have 5 transfers | 2, 5 have 6 transfers | 3, 14 have 7 transfers (Current seating has an absolute deviation of 0.74) 19 players: 4 18 7 11 6 | 19 9 12 3 1 | 15 13 17 2 8 | 14 10 5 16 (fixed in 1.5f) Rule 2 KO (player 1 shares a table with player 3): 66 (once), 3 (twice) (Current seating is: 56 once, 8 twice) Rule 3 KO. Absolute deviation is: 0,552631578947368 => 1, 2, 3, 4, 6, 7, 8, 9, 11, 12, 13, 15 have 10 VP | 5, 10, 14, 17, 18, 19 have 9 VP | 16 have 8 VP (Current seating has a better absolute deviation: 0,5) Rule 8 KO. Absolute deviation is: 0,648199445983379 => 1, 4, 6, 7, 11, 12, 13, 14, 15, 16, 17, 18, 19 have 5 transfers | 2, 9, 10 have 6 transfers | 3, 5, 8 have 7 transfers (Current seating has an absolute deviation of 0.69) 20 players: 9 4 12 16 11 | 14 18 7 2 6 | 19 10 17 13 1 | 20 15 5 8 3 Rule 2 KO (player 3 shares a table with player 5): 72 (once), 4 (twice) (Current seating is: 58 once, 11 twice) Rule 8 KO. Absolute deviation is: 0,72 => 1, 6, 7, 9, 11, 12, 14, 16, 17, 18, 19, 20 have 5 transfers | 2, 4, 10, 15 have 6 transfers | 3, 5, 8, 13 have 7 transfers (Current seating has an absolute deviation of 0.76) 21 players: 13 20 7 15 1 | 9 16 2 18 | 4 12 19 14 | 21 8 17 10 | 5 3 11 6 Rule 2 KO (player 3 shares a table with player 5): 66 (once), 1 (twice) (Current seating is: 58 once, 11 twice) Rule 3 KO. Absolute deviation is: 0,533333333333333 => 1 have 10 VP | 2, 3, 4, 5, 7, 13, 15, 20 have 9 VP | 6, 8, 9, 10, 11, 12, 14, 16, 17, 18, 19, 21 have 8 VP (Current seating has a better absolute deviation: 0.495) Rule 8 KO. Absolute deviation is: 0,258503401360544 => 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 16, 18, 19, 20, 21 have 5 transfers | 15 have 6 transfers | 17 have 7 transfers (Current seating has the same absolute deviation) 22 players: 18 21 16 19 1 | 22 17 12 2 6 | 9 13 7 15 | 4 8 10 11 | 14 5 20 3 Rule 2 KO (player 3 shares a table with player 5): 66 (once), 5 (twice) (Current seating is: 60 once, 8 twice) Rule 3 KO. Absolute deviation is: 0,472727272727273 => 1, 2, 6 have 10 VP | 3, 4, 5, 7, 8, 9, 10, 12, 16, 17, 18, 19, 21, 22 have 9 VP | 11, 13, 14, 15, 20 have 8 VP (Current seating has a better absolute deviation: 0,254) Rule 8 KO. Absolute deviation is: 0,446280991735537 => 1, 4, 6, 7, 8, 9, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22 have 5 transfers | 2, 5 have 6 transfers | 3, 10 have 7 transfers (Current seating has the same absolute deviation) 23 players: 23 18 21 11 16 | 19 22 17 12 6 | 9 13 7 20 1 | 14 4 2 8 | 15 5 10 3 Rule 2 KO (player 2 shares a table with player 4): 72 (once), 6 (twice) (Current seating is 76 once, 4 twice, which is better) Rule 3 KO. Absolute deviation is: 0,382608695652174 => 1, 6, 7, 9, 11, 12, 13 have 10 VP | 2, 3, 4, 5, 8, 10, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23 have 9 VP (Current seating has the same absolute deviation) Rule 8 KO. Absolute deviation is: 0,578449905482042 => 1, 2, 6, 7, 9, 11, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23 have 5 transfers | 3, 8, 10 have 7 transfers | 4, 5, 12 have 6 transfers (Current seating has an absolute deviation of 0.87) Note also that rule 7. "A player doesn't play in the same seat position, if possible." is violated in the current seating (player 3, table 1, position 3) 24 players: 19 23 17 21 16 | 24 18 22 11 6 | 14 20 12 1 7 | 4 9 2 13 8 | 15 10 5 3 Rule 2 KO. Rule 3 KO. Absolute deviation is: 0,466666666666667 => 1, 2, 4, 6, 7, 8, 9, 11, 12, 13, 14, 16, 17, 18, 19, 20 have 10 VP | 3, 5, 10, 15, 21, 22, 23, 24 have 9 VP (Current seating has the same absolute deviation) Rule 8 KO. Absolute deviation is: 0,666666666666667 => 1, 2, 4, 6, 11, 12, 14, 15, 16, 17, 18, 19, 21, 22, 23, 24 have 5 transfers | 3, 5, 8, 13 have 7 transfers | 7, 9, 10, 20 have 6 transfers (Current seating has an absolute deviation of 0.95) 25 players*: 10 14 22 18 1 | 15 19 2 23 6 | 20 24 7 3 11 | 25 4 12 8 16 | 5 9 17 13 21 (seating was done manually) Rule 2 OK. (Current seating is 84 once, 8 twice) Rule 8 KO. Absolute deviation is: 0,72 => 1, 2, 4, 6, 9, 11, 12, 15, 16, 17, 19, 21, 22, 23, 24 have 5 transfers | 3, 5, 8, 13, 18 have 7 transfers | 7, 10, 14, 20, 25 have 6 transfers (Current seating has an absolute deviation of 0.72) 26 players: 26 21 24 19 11 | 22 25 20 16 12 | 14 17 7 23 | 4 13 2 15 | 18 8 5 6 | 9 3 10 1 Rule 2 KO. Rule 3 KO. Absolute deviation is: 0,410256410256411 => 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 15, 18, 19, 20, 21, 23, 24, 25, 26 have 9 VP | 12, 13, 14, 16, 17, 22 have 8 VP (Current seating has an absolute deviation of 0.51) Rule 8 KO. Absolute deviation is: 0,390532544378698 => 1, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26 have 5 transfers | 2, 4 have 6 transfers | 3, 5 have 7 transfers (Current seating has an absolute deviation of 1.34) 27 players: 19 26 17 24 11 | 27 22 25 20 1 | 23 18 21 16 6 | 14 9 12 7 | 4 13 2 8 | 15 10 5 3 Rule 2 KO. Rule 3 KO. Absolute deviation is: 0,111111111111111 => 1, 6, 11 have 10 VP | 2, 3, 4, 5, 7, 8, 9, 10, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27 have 9 VP (Current seating has an absolute deviation of 0.40) Rule 8 KO. Absolute deviation is: 0,518518518518518 => 1, 2, 4, 6, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27 have 5 transfers | 3, 5, 8 have 7 transfers | 7, 9, 10 have 6 transfers (Current seating has an absolute deviation of 1.38) 29 players*: 24 28 22 26 21 | 29 23 27 16 11 | 19 14 17 6 1 | 4 18 12 7 2 | 25 9 20 13 8 | 15 10 5 3 Rule 2 KO. Rule 3 KO. Absolute deviation is: 0,425287356321839 => 1, 2, 4, 6, 7, 8, 9, 11, 12, 13, 14, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25 have 10 VP | 3, 5, 10, 15, 26, 27, 28, 29 have 9 VP (Current seating has an absolute deviation of 0.42) Rule 8 KO. Absolute deviation is: 0,677764565992866 => 1, 4, 6, 11, 12, 15, 16, 17, 18, 19, 21, 22, 23, 24, 25, 26, 27, 28, 29 have 5 transfers | 2, 7, 9, 10, 14 have 6 transfers | 3, 5, 8, 13, 20 have 7 transfers (Current seating has an absolute deviation of 1.33) 30 players*: 24 28 22 26 21 | 29 23 27 16 11 | 19 14 17 6 1 | 4 9 12 7 2 | 30 25 20 18 13 | 15 10 5 8 3 Rule 2 KO. Rule 8 KO. Absolute deviation is: 0,72 => 1, 4, 6, 11, 12, 15, 16, 17, 19, 21, 22, 23, 24, 26, 27, 28, 29, 30 have 5 transfers | 2, 7, 9, 10, 14, 25 have 6 transfers | 3, 5, 8, 13, 18, 20 have 7 transfers (Current seating has an absolute deviation of 1.52) 31 players: 23 26 29 25 28 | 27 30 21 24 17 | 31 22 19 20 18 | 14 8 12 16 | 4 13 7 11 | 9 5 2 6 | 15 3 10 1 Rule 2 KO. Rule 3 KO. Absolute deviation is: 0,16589861751152 => 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 have 9 VP | 16 have 8 VP (Current seating has an absolute deviation of 0.55) Rule 8 KO. Absolute deviation is: 0,468262226847034 => 1, 2, 3, 4, 6, 7, 8, 9, 11, 12, 13, 14, 15, 16, 20, 21, 22, 23, 24, 26, 27, 28, 29, 30, 31 have 5 transfers | 5, 17, 25 have 6 transfers | 10, 18, 19 have 7 transfers (Current seating has an absolute deviation of 1.32) Computing results for 30+ players in progress... * didn't perform an exhaustive search Prince of Paris, France Ratings Coordinator, Rules Director

Attachment Work_and.jpg not found

Seating for 17 players is interesting: the current archon has a better absolute deviation for VPs, and a worse one for transfers. And no, you can't have both Question is: would you favor the transfers so they are more equitably distributed, or the VPs (ie: how many times you sit at a 4-players table)? Prince of Paris, France Ratings Coordinator, Rules Director

> Ankha wrote: > Question is: would you favor the transfers so they are more equitably distributed, or the VPs (ie: how many times you sit at a 4-players table)? > I would rather the VPs be more equitably distributed than the transfers.

> BenPeal wrote: > > > Ankha wrote: > > Question is: would you favor the transfers so they are more equitably distributed, or the VPs (ie: how many times you sit at a 4-players table)? > > > I would rather the VPs be more equitably distributed than the transfers. > I guess you have to balance your data. Because going from 0.5 to 0.56 for the VP is no big deal but going from 0.93 to 0.31 for transfers is more important (IMO).

> Timo wrote: > > > BenPeal wrote: > > > > > Ankha wrote: > > > Question is: would you favor the transfers so they are more equitably distributed, or the VPs (ie: how many times you sit at a 4-players table)? > > > > > > > > I would rather the VPs be more equitably distributed than the transfers. > > > I guess you have to balance your data. > > Because going from 0.5 to 0.56 for the VP is no big deal but going from 0.93 to 0.31 for transfers is more important (IMO). > I'll do a test and keep only the best deviation for VPs (0.5) and see what we obtain as transfer deviation. Another thing is: would you like to play your 2 rounds on a 4 players table and have more transfers, or play a 5-players and a 4-players tables with "slightly" less transfers ("slightly" being the big unknown). Prince of Paris, France Ratings Coordinator, Rules Director

> Timo wrote: > > > BenPeal wrote: > > > > > Ankha wrote: > > > Question is: would you favor the transfers so they are more equitably distributed, or the VPs (ie: how many times you sit at a 4-players table)? > > > > > > > > I would rather the VPs be more equitably distributed than the transfers. > > > I guess you have to balance your data. > > Because going from 0.5 to 0.56 for the VP is no big deal but going from 0.93 to 0.31 for transfers is more important (IMO). > Except that the score of the game is based on Victory Points, not transfers.

> BenPeal wrote: > Except that the score of the game is based on Victory Points, not transfers. > Indeed it is, but getting stuff out one round earlier might make the difference between getting most/more (if any) Victory Points though... Ain't so simple but I kind trust that you guys are way ahead o me thinking these things out... Finnish Politics!

> Ankha wrote: > > 17 players: 13 16 7 1 6 | 17 3 11 2 | 4 8 15 10 | 5 12 9 14 > Rule 2 KO. > Rule 3 KO. Absolute deviation is: 0,558823529411765 => 1 have 10 VP | 2, 3, 4, 5, 6, 7, 13, 16 have 9 VP | 8, 9, 10, 11, 12, 14, 15, 17 have 8 VP > (Current seating has a better absolute deviation 0.5: 1, 2, 3, 4, 5, 6, 8, 11, 13, 17 have 9 VP | 7, 9, 10, 12, 14, 15, 16 have 8 VP) > Rule 8 KO. Absolute deviation is: 0.311418685121107 => 1, 3, 4, 5, 6, 7, 8, 10, 11, 12, 13, 14, 15, 16, 17 have 5 transfers | 2 have 6 transfers | 9 have 7 transfers > (Current seating has an absolute deviation of 0.93) > EDIT: THERE IS a seating with a VP absolute deviation of 0.5 and a better transfer absolute deviation: 11 8 6 10 7 | 9 4 14 1 | 5 3 15 2 | 12 17 13 16 3. Available VPs are equitably distributed. KO. Absolute deviation is: 0,5 => 1, 2, 3, 4, 5, 6, 7, 8, 10, 11 have 9 VP 9, 12, 13, 14, 15, 16, 17 have 8 VP 8. Starting transfers are equitably distributed. [NOAL] KO. Absolute deviation is: 0,816608996539792 => 11 have 3 transfers 6, 12, 14 have 4 transfers 1, 3, 4, 5, 8, 10, 15 have 5 transfers 2, 7, 9, 17 have 6 transfers 13, 16 have 7 transfers There's still one player with only 3 transfers which is lame. With the seatings that favor transfers: 1, 3, 4, 5, 6, 7, 8, 10, 11, 12, 13, 14, 15, 16, 17 have 5 transfers 2 have 6 transfers 9 have 7 transfers The VP distribution is: 1 have 10 VP 2, 3, 4, 5, 6, 7, 13, 16 have 9 VP 8, 9, 10, 11, 12, 14, 15, 17 have 8 VP Basically, we have one more player with 8 VPs, one more player with 10 VPs instead of two players with 9 VPs The player with 10 VP has the lowest transfers though (5). The player with the most transfers (7) has the lowest VPs (8). Thoughts? Prince of Paris, France Ratings Coordinator, Rules Director

Seems fair to me.

Hmpf. I've rewritten a faster algorithm, and discovered that there was a better seating for 17 players: 9 16 11 6 10 | 13 3 7 1 | 4 8 2 14 | 5 12 17 15 Rule 2 KO. Rule 3 KO. Absolute deviation is: 0,5 => 1, 2, 3, 4, 5, 6, 9, 10, 11, 16 have 9 VP | 7, 8, 12, 13, 14, 15, 17 have 8 VP (Current seating has a the same absolute deviation) Rule 8 KO. Absolute deviation is: 0,311418685121107 => 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 16 have 5 transfers | 15 have 6 transfers | 17 have 7 transfers (Current seating has an absolute deviation of 0.93) So it's the best of the two worlds. Prince of Paris, France Ratings Coordinator, Rules Director

I was looking for hese famous "rules" and was amazed to find that i was not allowed to have a player be twice in the 5th seat but nothing prevented a player to be twice in the 1st seat. I don't understand he reason behind this rule. What I understand in here that LSJ considered (back in 2002) that 5th was the worst place. I disagree ! IMO, 1st is the worst case scenario 1 alone transfer is meaningless in most deck because you can't influence out a meaningfull vampire with only 1 transfer (I mean except in dedicated decks like LOP or DBR). So why forbidding a player being 5th twice while allowing a player to be 1st twice. And in the case of a 5 players tournament, we have player 2 being 1st twice and 2nd the 3rd time (for a cumulative total of 4 transfers in his firt turns) while player 4 have 12 transfers being 4th twice and 5th the 3rd time !! I found this to be strange... So what is your view of it ?

> Timo wrote: > I was looking for hese famous "rules" and was amazed to find that i was not allowed to have a player be twice in the 5th seat but nothing prevented a player to be twice in the 1st seat. > Rule 7 prevents it: 7. A player doesn't play in the same seat position, if possible. Interesting link anyway (I understand at least what "NOAL" means, thought it was something like "no alignment" ) Anyway, since the goal is to distribute equitably transfers, you can't be first then second for instance. Being 5th is worse than being 4th. But I don't know if it's worse than being first. I remember lost game because I was always one turn late vs my predator. Prince of Paris, France Ratings Coordinator, Rules Director

Ok, it is debatable if 5th is better or worse than 1st. But actually, it IS debatable... Which also means that there is a criteria which prevent one of those starting place to be duplicated while the other can be (and is). Why should it stay like that ? Is it possible to review these "rules" ? (and maybe update them).

> Timo wrote: > Ok, it is debatable if 5th is better or worse than 1st. > > But actually, it IS debatable... Which also means that there is a criteria which prevent one of those starting place to be duplicated while the other can be (and is). > No. If possible, seating positions can't be duplicated. Or I don't understand what you're saying. > Timo wrote: > Why should it stay like that ? > Moot because your premise is wrong. Prince of Paris, France Ratings Coordinator, Rules Director

The rules I found in the link above are these ones : > 1. No pair of players repeat their predator-prey relationship. This is mandatory, by the VEKN rules. > 2. No pair of players share a table through all three rounds, when possible. > > 3. Available VPs are equitably distributed. > 4. No pair of players share a table more often than necessary. > 5. A player doesn't sit in the fifth seat more than once. > 6. No pair of players repeat the same relative position[*], when possible. > 7. A player doesn't play in the same seat position, if possible. > > 8. Starting transfers are equitably distributed. [NOAL] > 9. No pair of players repeat the same relative position group[^], when possible. > And we have rule 5 who is considered more important than rule 7 which have an "if possible" in it. I was speacking about that.

> Timo wrote: > The rules I found in the link above are these ones : > > > 5. A player doesn't sit in the fifth seat more than once. > > 7. A player doesn't play in the same seat position, if possible. > > > Currently (for seatings up to 23), rule 7 has never been violated: it was always possible to seat the player in another position. > This is also enforced by the fact that in order to have the best distribution of transfers, a player can't seat twice at the same position. > > So, there's no need right now to change rule 7 or 5: rule 7 is a superset of rule 5 and the "if possible" clause has never been broken. > Prince of Paris, France Ratings Coordinator, Rules Director

You are in your hunt for the best seating for 2R+F tournament. But I was speaking of the seatings for 3R+F tournament where for 5 players tournaments, we have player 2 who play 2nd then 1st and 1st. Without looking for extreme cases like 5 players tournaments, in 10 players tournaments, players 2 and 7 play twice in the second place and 3 and 8 twice in the 4th place. EDIT : and in 15-players tournament, 3 and 8 play twice in the 3rd position (I will not check furher because i is obviously less likely to happen and I do that "manually" checking the tables in the archon (1.5c which I downloaded yesterday here in vekn.net)

> Timo wrote: > You are in your hunt for the best seating for 2R+F tournament. > > But I was speaking of the seatings for 3R+F tournament where for 5 players tournaments, we have player 2 who play 2nd then 1st and 1st. > > Without looking for extreme cases like 5 players tournaments, in 10 players tournaments, players 2 and 7 play twice in the second place and 3 and 8 twice in the 4th place. > You're speaking about the current seatings. It doesn't mean there isn't a better seating for which rule 7 applies in full. We must first search better seatings that enforce the current rules before thinking of changing them. Prince of Paris, France Ratings Coordinator, Rules Director

Why that ? If the rules are to be changed, I guess it is for the best to do it BEFORE looking for better seatings. I mean : if after review, one decides to change the rules (I am no sure whose role it is to change them), it would be logical to check afterward the current seatings to look for better ones... What I mean is that the rules which FORBID to have a player playing 5th in 2 different rounds could potentially prevent a better VP/transfer repartition while having no good effect. (I state here what is my opinion that forbiding the repetition of the 5th place while trying to share evenly the transfer and to avoid if possible the repetition of any place is redundant while potentially bad)

> Timo wrote: > Why that ? > > What I mean is that the rules which FORBID to have a player playing 5th in 2 different rounds could potentially prevent a better VP/transfer repartition while having no good effect. > Because the more constraints we have, the less possibilities we have. And since possibilities are exponentials, having the tighest constraints is essential to achieve computation. The rule "COULD potentially prevent" (emphasis mine) some better results. But maybe it doesn't. On 2R it doesn't change anything, I haven't studied 3R yet. If you have some spare time, try to find one seating where the rule prevents a better seating. If it does, okay we may think about dropping it. If it doesn't, it's a easy way to reduce the number of possible seatings. Prince of Paris, France Ratings Coordinator, Rules Director

It is only a constraint of computing time ? So you will gain some time when checking your conditions and so deleting this rule which is redundant with the rule 7 every time when you have reach a number of player big enough to avoid infringing rule 7 (always for 2R+F as you have pointed out). And for the few cases where you have to add some odd seatings which would have been already eliminated, I guess the additional compuing time is meaningless compare to the size of the data no ?

> Timo wrote: > It is only a constraint of computing time ? > > So you will gain some time when checking your conditions and so deleting this rule which is redundant with the rule 7 every time when you have reach a number of player big enough to avoid infringing rule 7 (always for 2R+F as you have pointed out). > > And for the few cases where you have to add some odd seatings which would have been already eliminated, I guess the additional compuing time is meaningless compare to the size of the data no ? > If you inforce strictly rule 7 (7. A player doesn't play in the same seat position, if possible.), it cuts branches in the tree of possibilities. For instance, when I must seat player 4 of a 10-players table, it removes 2 possibilities for 4. Let's say I have 5,6,7,8,9 and 10 to seat afterwards, I'm saving 2 * (5 * 4 * 3 * 2 * 1) possibilities (because there would be 5 possibilities for 5, 4 possibilities for 6, 3 for 7 etc.). A little less in truth because 5, 6, 7... also have constraints, but it's really a huge save. Prince of Paris, France Ratings Coordinator, Rules Director

I'm done with 2R+F computations for up to 31 players. Some computations couldn't have been done exhaustively (the more 5-players tables there is, the more exponential it becomes). I'm starting to work on 3R+F. Prince of Paris, France Ratings Coordinator, Rules Director

Last tournament we had (10 plrs with 3R+F) two players were on the same table the whole game. Apparently Marko's deck shut down Tom's pretty well in all of those so he wasn't exactly happy about it. vekn.net/forum/9-event-reports-and-twd/69254 Used the latest Archon (1,5c) for the first time and I don't reacll if I ever encountered that before with the previous versions. Not really optimal if you play all three games in the same table as one other player... :/ Finnish Politics!

> Lönkka wrote: > Last tournament we had (10 plrs with 3R+F) two players were on the same table the whole game. Apparently Marko's deck shut down Tom's pretty well in all of those so he wasn't exactly happy about it. > > vekn.net/forum/9-event-reports-and-twd/69254 > > Used the latest Archon (1,5c) for the first time and I don't reacll if I ever encountered that before with the previous versions. > > Not really optimal if you play all three games in the same table as one other player... :/ > Archon 1.5c contains the historical seating. The new seatings for 2R+F will be in the next archon. I haven't started to work on 3R+F yet, but with 10 players, there is a high possiblity that you are obliged to have 2 players share the same table the whole 3 rounds. Prince of Paris, France Ratings Coordinator, Rules Director

Haven't checked my old 10 player Archons if the same thing happened there or was this just an isolated case (probably not). Finnish Politics!

I do not think it is possible with 2 tables and 10 players to avoid having at least 1 pair that plays all three games with each other. If you only have two tables, each player has to play on one of them at least twice. The only way to avoid this is more players then 10 so you get more then 2 tables. David Tatu V:EKN Scribe

It's impossible indeed. R1: [1 2 3 4 5] and [6 7 8 9 10] > Consider player 1. On round 2, let's say he must be sitting with noone of the first table, so : > > R2: [1 6 7 8 9] and the rest: [2 3 4 5 10] > > Now consider player 6: he must sit with players others than 7, 8 and 9 so: > > R3: [6 10 3 4 5] which gives as the last table [1 2 7 8 9] > > So 7, 8 and 9 have shared 3 games. > > Or let's say he sits with a player from the first round (eg. 2): > > R2: [1 2 6 7 8] and [3 4 5 9 10] > > Considering again player 6: > > R3: [6 9 10 1 3] and [2 4 5 7 8] > > So 7 and 8 have shared 3 games > > Or let's say he sits with two players from the first round (eg. 2 and 3): > > R2: [1 2 3 6 7 ] and [4 5 8 9 10] > > Considering again player 6: > > R3: [6 8 9 10 1] and [2 3 4 5 7] (8 and 9, 2 and 3 have shared 3 games) > or [6 8 9 10 2] and [1 3 4 5 7] (8 and 9, 3 and 4 have shared 3 games) > or [6 8 9 10 3] and [1 2 4 5 7] (8 and 9, 4 and 5 have shared 3 games) > or [6 8 9 10 4] and [1 2 3 5 7] > etc. > > Or let's say he sits with three players from the first round (eg. 2, 3 and 4): > > R2: [1 2 3 4 6 ] and [5 7 8 9 10] > > Considering again player 6: > > R3: [6 7 8 9 10] and [1 2 3 4 5] (7, 8, 9 and 10 shared 3 games) > etc. > Prince of Paris, France Ratings Coordinator, Rules Director

Is it possible to change around seating order in a way that those two that has to sit on same table in all three games would only once be in predator/prey relations? Because right now the three games goes like this: game 1: A is prey of B game 2: B is prey of A game 3: A is grand prey of B I think it would be better as: game 1: A is prey of B game 2: B is grand prey of A game 3: A is grand prey of B ...if possible. Marko, Prince of Helsinki

> a-e wrote: > Is it possible to change around seating order in a way that those two that has to sit on same table in all three games would only once be in predator/prey relations? > > Because right now the three games goes like this: > game 1: A is prey of B > game 2: B is prey of A > game 3: A is grand prey of B > > I think it would be better as: > game 1: A is prey of B > game 2: B is grand prey of A > game 3: A is grand prey of B > > ...if possible. > It's probably not possible. Prince of Paris, France Ratings Coordinator, Rules Director

This is current optimal seating for 10 player, players 8 and 9 are in same table in all three games. Round 1 1,2,3,4,5 6,7,8,9,10 Round 2 10,1,9,8,2 5,6,4,3,7 Round 3 4,7,10,3,1 9,2,5,8,6 If on round 2 players 1 and 9 would change places, would it break it? Marko, Prince of Helsinki

> a-e wrote: > This is current optimal seating for 10 player, players 8 and 9 are in same table in all three games. > > Round 1 > 1,2,3,4,5 > 6,7,8,9,10 > Round 2 > 10,1,9,8,2 > 5,6,4,3,7 > Round 3 > 4,7,10,3,1 > 9,2,5,8,6 > > If on round 2 players 1 and 9 would change places, would it break it? > Just check by yourself: 9 and 10 would be prey/predator on table 1, and predator/prey on table 2. Prince of Paris, France Ratings Coordinator, Rules Director

> Ankha wrote: > > > a-e wrote: > > This is current optimal seating for 10 player, players 8 and 9 are in same table in all three games. > > > > Round 1 > > 1,2,3,4,5 > > 6,7,8,9,10 > > Round 2 > > 10,1,9,8,2 > > 5,6,4,3,7 > > Round 3 > > 4,7,10,3,1 > > 9,2,5,8,6 > > > > If on round 2 players 1 and 9 would change places, would it break it? > > > Just check by yourself: 9 and 10 would be prey/predator on table 1, and predator/prey on table 2. > Yes. But the question is, is it better or worse than what we have now? 8 and 9 being predator/prey on first round, then prey/predator on second, same what 9 and 10 would be, and then predator/grand prey on third round. 8 and 9 would still play three games in same table, but only once next to each other. 9 and 10 would play twice next to each other but they don't share table three times. Marko, Prince of Helsinki

Exhaustive search done for 24 players 2R+F, more than a week of computation using distributed computers. Prince of Paris, France Ratings Coordinator, Rules Director

Seatings are being reevaluated as rule 2 was considered less important than rules 3 and 8. The following seatings have already been updated: 12, 13, 14, 15, 17. Prince of Paris, France Ratings Coordinator, Rules Director

> Ankha wrote: > Seatings are being reevaluated as rule 2 was considered less important than rules 3 and 8. > The following seatings have already been updated: > 12, 13, 14, 15, 17. > 18 has been updated. Prince of Paris, France Ratings Coordinator, Rules Director

19 has been updated. Prince of Paris, France Ratings Coordinator, Rules Director

20, 21 and 22 have been updated. Prince of Paris, France Ratings Coordinator, Rules Director

There was a flaw in the validation of the new results, and some manual mistakes when reporting the results. As a consequence, some of the results displayed here are wrong, and I won't continue to maintain this page: the new results will always be available in the latest version of the archon. From now on, I strongly recommand not to use the version 1.5e but version 1.5f : Excel: www.vekn.net/images/stories/downloads/thearchon1.5f.xlsx LibreOffice: www.vekn.net/images/stories/downloads/thearchon1.5f.ods Prince of Paris, France Ratings Coordinator, Rules Director