Forty-third KGS Computer Go Tournament

Sunday October 12th 2008

These results also appear on official KGS pages: Formal Division, Open Division which link to the game records.


 Formal divisionOpen division
format6-round Swiss9-round Swiss
board size13×139×9
time28 minutes plus 25/20s18 minutes plus 25/20s


The first round started at 08:00 UTC for the Formal and 08:05 for the Open division.


As usual, the tournament was held in two divisions, Formal and Open, with more restrictive entry conditions for the Formal division.

Formal Division   13×13


Open Division   9×9


In the past these events have always used "absolute time", so as to allow the tournament to run on schedule. This month, for the first time, I used Canadian overtime of 25 stones in 20 seconds. The purpose is to ensure that a player that reaches the end of the sensible part of the game within the main time allocation, cannot be made to lose on time by an opponent that plays on pointlessly.
       My intention was to use Canadian overtime of 20 stones in 20 seconds. Because of a bug in the tournament setup program, I can't do this. But 25 stones in 20 seconds is good enough, and will be used in future.
       The change of the timing system had no effect on this tournament. Only two games were won on time, and in both this was because one of the players has ceased to move at all. This is not surprising as this tournament used small boards with ample main time allocations.

A bot new to these events was registered for the Open division, Juho Pennanen's Rover. Unfortunately its registration email was sent only minutes before play started, and I did not see it (I can't swear that I didn't actually receive it) until too late.
       I prefer to receive registration emails at least half an hour before play starts. In principle I can accept them until two minutes before play starts, but I am often fairly busy in the last few minutes, and so this is not reliable. If you do decide to enter in the last few minutes, it may be better to say so in the chat of the Computer Go room.

Again no entries were received from MoGo or from CrazyStone. I again accepted an entry from CzechBot, a build of MoGo by Petr Baudiš, on the understanding that I would remove it if MoGo's authors entered their own version.

Formal division

HouseBot vs AyaMC
Position after move 143

In round 1, HouseBot and AyaMC reached the position shown to the right SGF. HouseBot (White) played in the corner putting its two stones in atari, AyaMC captured them, and HouseBot tried to recapture. This move was rejected by the server because it violates the positional superko rule, and HouseBot eventually timed out. Meanwhile valkyria19 resigned against CzechBot.

In round 2, AyaMC resigned against CzechBot, and valkria19 resigned against ManyFaces1.

In round 3, AyaMC resigned against ManyFaces1, and HouseBot resigned against CzechBot.

CzechBot vs ManyFaces1
Moves 87-91
In round 4, HouseBot resigned against valkria19. CzechBot and ManyFaces1 had an interesting game, in which ManyFaces1 killed CzechBot's group with the sequence shown to the left. CzechBot eventually resigned. SGF

In round 5, AyaMC resigned against valkyria19, and HouseBot resigned against ManyFaces1.

So after five rounds, ManyFaces1 had won all its games, CzechBot had lost only to ManyFaces1, valkyria19 had lost only to ManyFaces1 and CzechBot, and AyaBot had lost only to those three. So it looked as if we could have a simple results list, with every player winning all its games against those ranked below it.

This all ended in round 6, which reversed two previous results, with AyaMC beating valkyria19, and CzechBot beating ManyFaces1 in an interesting game SGF. Thus at the end, CzechBot and ManyFaces1 had beaten each other once each and beaten all the other players, while AyaMC and valkyria19 had beaten each other once each and each beaten HouseBot twice.

Open division

In round 1, break resigned early in a lost position against ManyFaces2. This game is shown in the KGS records as forfeited, but I know of no reason for this, I recall it as a normal resignation. HBotSVN resigned against AyaMC2.

In round 2, break resigned against HBotSVN, and AyaMC2 resigned against ManyFaces2.

In round 3, break resigned against AyaMC2, and HBotSVN resigned against ManyFaces2.

In round 4, AyaMC2 resigned against ManyFaces2, and break resigned against HBotSVN.

In round 5, break resigned against AyaMC2, and HBotSVN resigned against ManyFaces2.

In round 6, break vanished after two moves of its game against manyFaces2, and eventually timed out. HBotSVN resigned against AyaMC2.

In round 7, break reappeared in time for its game against HBotSVN, but had to resign. AyaMC2 resigned against ManyFaces2.

ManyFaces2 vs HBotSVN
Position at game end.
In round 8, break resigned against AyaMC2. ManyFaces2 and HBotSVN achieved a seki, which was still on the board at the game end, as shown to the right. I was apprehensive about what might happen next, but HBotSVN (Black) resigned, suggesting that it had recognised the position as seki SGF. However, its author Jason House has told me that it has no concept of seki; it resigned simply because it knew it had no way of winning.

In round 9, break resigned against HBotSVN, and ManyFaces2 resigned against AyaBot, reversing three earlier results. Overall, ManyFaces2 won all its games except this one with AyaBot; AyaBot won all its games except three against ManyFaces2, and HouseBot won all its games against break.


Processor numbers, power, etc.

Aya, running on Xeon X5355 2.66GHz 2CPU (4 cores on 1CPU, so all 8 cores)
Aya, running on AthlonX2 2.3GHz 1CPU (2 cores on 1CPU)
break, probably running on a single processor Intel(R) Celeron(R), 1.7Ghz
MoGo, running on 2 cores of a four-way Intel(R) Xeon(R) CPU 5130 @ 2GHz
AMD Athlon(tm) 64 X2 Dual Core Processor 5000+
AMD Athlon(tm) 64 X2 Dual Core Processor 5000+
Many Faces of Go, running on a 32-core Xeon 3.2GHz: 4 nodes, 8 cores per node, connected by a 40Gbps network
Many Faces of Go, running on a 32-core Xeon 3.2GHz: 4 nodes, 8 cores per node, connected by a 40Gbps network
valkyria, running on a single processor 3Ghz P4