[Trhm-Go] Turneringer
Robert Biegler
Robert.Biegler at svt.ntnu.no
Fri Sep 25 16:40:21 CEST 2009
-- Turneringsform. Det hadde vært best med noe som gir alle spillerne
like gode sjanser til å vinne. MacMahon fungerer vel til
formålet, men Robert foreslo en mer gamblerinspirert variant.
I suggest adapting an idea by Hasler, proposed for sailing. When people race each other in boats that are not identical, you need a handicapping system. Hasler proposed ding away with attempts to predict boats' speeds by measuring them, and let their owners do the estimation. To remove incentives to claim low speeds to gain a handicapping advantage, Hasler proposed the other sailors would decide how many points to bet on beating someone at the handicap he or she claims. I think Hasler's idea could be transferred to Go as follows, if everyone plays everyone else:
Each player posts the rank under which they want to play. We have a table that says how big a rank difference corresponds to how many handicap stones and how much komi (so we can accommodate differences in playing strength greater than 9 handicap stones). Each player has a fixed number of points that must be allocated to the games, with a minimum of 1 point per game. Let's say there are 10 players, so each plays 9 games. Say we allocate a total of 36 points per player, an average of 4 points per game per player. If Sverre gave himself a rank equal to mine, I would say I have little chance of beating him in an even game, and would give that game the minimum number of points, namely 1. If he declared himself 8 ranks above me, I would put on that game all the points I don't have to allocate to other games. I have 8 other games, each getting a minimum of 1 point, that leaves me 28 points to bet on my game with Sverre. The winner of a game gets the points both players bet on the game. The winner of the tournament is the player with the most points at the end.
If we don't have all players playing everyone else, it may be feasible to set up the initial betting as just described and just ignore the missing games.
The advantage of the system is that if the players know each other's strength, there is no incentive to declare a lower rank than justified, because that would just mean the other players bet minimum points on those games and the player would gain little from all those victories. If the players don't know each other, there is no basis for deciding whether the rank someone claims deviates from their real strength, and the betting doesn't work.
Regards
Robert
More information about the Trhm-Go
mailing list