Sixty-first KGS Computer Go Tournament

Sunday July 4th 2010

These results also appear on an official KGS page which links to the records of all the games.


format7-round Swiss
board size19×19
time29 minutes plus 25/30s


The first round started at 16:00 UTC.

Result table



In round 1, EricaBot got a lost position against hcBot. Its operator and author, Shih-Chieh Huang, asked me if he was allowed to resign for it. I told him he was. Normally, any kind of intervention with the program during the course of a game is forbidden (though settings may be changed between rounds). But resignation is always allowed, and I told him so. He resigned its game.
       There are several ways you can in effect cause a bot to resign. One is to disconnect it from the server, and leave it to lose on time – this will be shown in the game record as a loss on time. Another is to disconnect it from the server, log in to its account, and resign for it. But if you do this, the server will think you are trying to cheat by logging in to a bot's account from a human client, and show if as a loss by forfeit. The best way is to arrange to be able to tell your program that it must resign.

WeakBot50k vs Orego
Move 87.

In round 2 Orego captured some of WeakBot50k's stones in a ladder SGF (WeakBot50k always pulls its stones out of atari if it can). It then allowed WeakBot50k (which is also keen on capturing stones) to recapture all the laddering stones, one by one, while it secured the rest of the board, as seen to the right.

In round 4, AyaMC beat Zen19, in what turned out to be the decisive game of the tournament SGF.

Orego vs CzechBot
Moves 225-229.
228 captures 225,
229 captures 226.

Also in round 4, CzechBot had a won position against Orego SGF when the sequence shown to the left was played. Orego wanted to play 230 to capture 227, but this was disallowed by the superko rule. When it was unable to make the move that it wanted, Orego stopped playing, and eventually lost on time. The is the first triple-ko I have seen in any computer Go tournament.

At this point there were four players all on three wins from four games. AyaMC had lost only to pachi, pachi had lost only to Zen19, and CzechBot and Zen19 had lost only to AyaMC.

Orego was late for its round 5 game with SimpleBot – I assume it had needed to be restarted by its operator.
       Also in round 5, the game between Zen19 and CzechBot SGF had a confusing ko fight from moves 192-207. I suspect that CzechBot could have played better here, and if it had, it might have won the ko fight and the game.

Now AyaMC, pachi and Zen19 were still tied on four wins from five games.

After an unremarkable round 6, AyaMC, pachi and Zen19 were still tied, now on five wins from six games.

In the final round 7, AyaMC beat Orego, and Zen19 beat PNUGo. But pachi lost to CzechBot, leaving only AyaMC and Zen19 on six wins. AyaMC was ahead on the SOS (Sum of Opponents' Scores) tiebreak, winning the tournament. I was pleased to see that AyaMC no longer times out after the game is properly over, as it did several times last month.

Details of processor numbers, power, etc.

Aya, running on i7 980X 3.3GHz 6cores
break, probably running on a single processor Intel(R) Celeron(R), 1.7Ghz
MoGo, running on double-core AMD Athlon(tm) 64 X2 Dual Core Processor 4800+ (2.5GHz).
Erica, running on Xoen 8 cores, 2.26 GHz.
an unofficial modified version of Fuego, running on Core i7 980X (6 cores at 3.3GHz), 12 threads
running on Linux, 2GiB RAM, Intel(R) Celeron(R) M CPU 530 @ 1.73GHz
Orego, unspecified but probably running on 2 x 3 GHz Dual-Core Intel Xeon
pachi, running on 36 threads mixed i7/core2 on 7 machines
GNU Go, unspecified platform
running on one processor of a 4GiB RAM, AMD Athlon(tm) 64 X2 Dual Core Processor 4000+, shared with WeakBot50k
running on Linux, 2GiB RAM, Intel(R) Celeron(R) M CPU 530 @ 1.73GHz
running on a Mac Pro 8 core, Xeon 2.26GHz.