Eighty-eighth KGS Computer Go Tournament

Sunday December 2nd, 2012

These results also appear on an official KGS page.


format18-round Swiss
board size9×9
time9 minutes plus 10/30s


The first round started at 16:00 UTC.

Result table

Crazy stv Fuego coldm AyaMC ManyF pachi Orego
B03R J8 B01140 W01622 W11R B110R W113R W16R B114R W117R W14R B19R W122 B1154 B05R W118R B17R W112R 13½166
2stv W13R J8 W11140 B11622
B07R W09R B017R B12R W110R B05R W118R W142 J13 W1146 W16R B112R B11R W115R 13164½109½
3Fuego9 B01R W010R B013R W17R B19R W117R
W15R B012R W015R W02R B18R W164 B1112 W1186 B14R W116R W13R B114R 1216594
4coldmilk9 B06R W014R B017R W02R B010R B05R W112R B115R
W11R B113R W116R W03R B1810 W17R B19R W14R B111R W118R 1115670
5AyaMC B04R W09R W15R B018R B12R W08R B01R W013R B016R
W07R B012R B13R W111R J14 W115R B16R W110R B117R 15147¾
6ManyFaces1 B022 W0154 B042 J13 B0146 B064 W0112 B0186 B13R W0810 B17R W112R
W11R B01017 W0176 B15R W19R B116R 160½
7pachi W15R B018R B06R W012R W04R B016R B07R W09R W03R B011R J14 B015R B01R W11017 B1176
B12R W18R B113R 155½
8Orego12 W07R B012R W01R B015R B03R W014R B04R W011R B018R W06R B010R W017R W05R B09R W016R W02R B08R W013R

In the table above,
   0 is a loss
   1 is a win
   J is jigo
   left superscript is the player's colour
   right superscript is the round in which the game was played
   a subscript shows how the result was determined:
      R for resignation
      T for time
      F for forfeit
      a number for the points difference after counting.
All the 0s, 1s and Js are links to the game record.

In the "SOS" and "SoDOS" columns, where a cell contains two numbers, the number in the larger font is calculated correctly, and the number in the smaller font is calculated by the KGS software and appears on the KGS results page.

Eight players registered for the tournament.


stv vs CrazyStone
Move 45

In round 3, the only two undefeated players, stv and CrazyStone, played each other. Stv played the first half of the game faster than CrazyStone: at move 25, stv had used two minutes while CrazyStone had used four. I realised when I saw CrazyStone's move 45 (shown to the right) that it had no hope of winning. This move is sente, but if answered in the obvious way, makes Black's position worse in all respects. MC bots make such moves only when their position is hopeless.

In round 4 stv beat ManyFaces1 in a difficult game with many kos.

AyaMC vs stv
Final move

In round 5 stv lost its first game, to AyaMC, by resigning (as Black) in the position shown to the left. This resignation is, I believe, correct. Black must connect, and White is then able to play at a9, starting a ko. White will need to play the first ko threat, and can do so at h4. Black then has no ko threat and so will lose the ko.

AyaMC vs ManyFaces1
After consecutive passes

In round 7, ManyFaces1 (Black) and AyaMC both passed in the position shown to the right, and then disputed the status of the "bent four in the corner" position in the top left. With good play this is alive for White, as White has a ko threat at g7 and Black has no ko threat. If it had been played out correctly the result would have been jigo. But AyaMC failed to save these stones, and lost the game.

In round 8 there was a jigo between CrazyStone and stv. In the other three games between these players, stv beat CrazyStone, the winner of the tournament. However third-placed Fuego9 won all its three games against stv, and lost all its three games against CrazyStone.

pachi vs ManyFaces1
After consecutive passes

In round 10, ManyFaces1 (Black) and then pachi passed in the position shown to the left. They disputed the status of the stone at b1, so play resumed, with White's final pass being undone. White passed again. So there had again been two consecutive passes, the clean-up phase was over, and the stone on b1 was scored as alive. In fact White would have won anyway, with correct scoring. But my impression is that the KGS software mis-handled the clean-up phase, both the passes should have been undone.

ManyFaces1 vs Fuego9
After consecutive passes

In round 11, Fuego (Black) and then ManyFaces1 passed in the position shown to the right. They then disagreed about the staus of the large white group in the top left, and entered the clean-up phase. In this, White passed throughout, while Black captured and connected the ko at a1 and then made some meaningless moves. So the game was scored with the large white group, and the three black stones in the bottom right, alive. The result was a win for Black, as it would have been with correct play.

stv vs Orego12
Moves 22-34
22 is at 33
26 is at 31
30 is at 33

In round 15, stv successfully executed a two-stone edge squeeze against Orego12, as shown to the left. I don't know if this impresses strong players. It impresses me because this squeeze is at the limit of my reading ability, even though I have trained myself to recognise opportunities for it.

Annual points

Players receive points for the 2012 Annual KGS Bot Championship as follows:

Crazy Stone8

Details of processor numbers, power, etc.

Aya, running all 12 threads on an i7-980X 3.3GHz 6cores +HT.
coldmilk, running on a 16-core Xeon, 2.9GHz.
Crazy Stone, running on a 24-core PC.
Fuego, running on a 2 core, 24 thread, Intel(R) Xeon(R) CPU, X5670 @ 2.93GHz.
Many Faces of Go, running on an Amazon EC2 Cluster Compute Eight Extra Large Instance, 2 x Intel Xeon E5-2670, using 16 cores, 16 threads, and 48 GB of memory.
Orego, running on one node of our custom Linux cluster Fido. The node has two AMD Six Core Opteron 2427 2.2 GHz processors (12 cores total), 8 GB RAM, Centos Linux
pachi, running on 2x Opteron 6134 (15 threads) with 64GiB RAM.
Steenvreter, probably running 46 threads each at 2.2 GHz, on a system whose use was generously provided by the Maastricht games and AI group.