Seventy-fifth KGS Computer Go Tournament

Sunday September 4th, 2011

These results also appear on an official KGS page.


format7-round Swiss
board size19×19
time29 minutes plus 10/30s


The first round started at 16:00 UTC.

Result table

Zen19 pachi stv AyaMC Czech gomor Simpl Orego PueGo oakfo
B04½ B11R B15R W17R B13R W16256½ B12R 62620Winner
2pachi2 W14½
W07R W11R W13172½ W15R B16R W12R 62419
3stv W01R B17R
B03R B14R W16R W15130½ B12R 52716
4AyaMC W05R B01R W13R
W12R W1736½ W16R W14R 52614
5CzechBot B07R W04R B02R
W15R B0165½ B13R W16R 3257
6gomorra3 W03R B06R B05R
B02T B14R W17R 3204
7SimpleBot B06256½ B03172½ B05130½ B0736½ W12T
B0492½ W11T 2295
8Orego12 W02R B05R B06R W1165½ W1492½
B07R B03R 2265
9PueGo W06R W02R W03R W04R B01T W17R
W15R 2234
10oakfoam B02R B04R B06R B07R W13R B05R

In the table above,
   0 is a loss
   1 is a win
   J is jigo
   a superscript is the round in which a 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.

The numbers in the table do not add up as you might expect. This is because they include a point for a win against WeakBot50k, and a point for a bye.

Eleven players entered, including WeakBot50k. My intention was to remove WeakBot50k if there were an odd number of registered players, but shortly before the tournament started, CzechBot had not appeared. I removed WeakBot50k soon after I saw CzechBot, but by then it had already been included in the draw for round 1. I removed it before the draw for round 2.


In round 1, stv appeared to get a good position aganist Zen19, but Zen19 outplayed it and won. Oakfoam misread a ladder against WeakBot50k, but won anyway. PueGo lost on time in a won position against SimpleBot. CzechBot obtained a won position against Orego12, but when it had only 3 seconds left on its clock, it started passing (it does not understand overtime), and lost.

pachi2 vs oakfoam
Move 170

In round 2, pachi2 played the move shown to the right, in its game with oakfoam. Three possible reactions from an observer are "This move is really stupid, even a 20-kyu can see that it achieves nothing, the player can't really be 3-dan", "if pachi2 is 3d and can make such stupid moves, how much stronger would it be if it learned not to make them?", and "this move wins the game as effectively as any other, we should not criticise it". I believe that only the third is correct, and I try to educate observers who react in the first two ways.

PueGo vs stv
around move 270
    Also in round 2, Gomorra3 lost on time in a won position against SimpleBot, CzechBot mishandled its time and lost to AyaMC, and PueGo and stv both correctly handled the seki shown to the left. Gomorra3's operator found that three of its eight nodes were hanging, and reduced it to playing on four nodes for round 3. He switched it back to eight working nodes after its round 3 game was over.

During round 3, CzechBot's operator found that its computer was using time to execute some unknown process, which he terminated.

Zen19 vs pachi2
move 49

In round 4, Zen19 played the move shown to the left, against pachi2. The British champion Matthew Macfadyen 6-dan happened to be watching – his opinion was that while this move looks eccentric, it may be the best way to settle the black hoshi stone.
    Pachi2 eventually won this game. It was Zen19's only loss in the tournament.

Zen19 vs CzechBot
Moves 71-77

During round 7, Zen19 appeared to miss a snapback in its game with CzechBot, as shown to the right. Zen19 still won the game.

pachi2 vs stv
Moves 48-93
Also in round 7, pachi2 mis-read a ladder against stv, as shown to the left, and resigned soon afterwards. It was pachi2's only loss in the tournament.

Annual points

Players receive points for the 2011 Annual KGS Bot Championship as follows:



Details of processor numbers, power, etc.

Aya, running on 6 cores of an i980X 3.3GHz
a development version of MoGo current as of 2010-10-06, running on an 8-thread i7 920, 6GiB RAM
Gomorra, running on a cluster of 8 nodes with 12 cores each, thus 96 cores. Cores are running at 2.67 Ghz. Nodes are connected through 4xSDR InfiniBand.
Oakfoam, running on a quad-core Intel Xeon E5520 with hyperthreading (so 8 threads).
Orego, probably running on one of the five nodes of a custom Linux cluster built by PSSC Labs: the node has two AMD Six Core Dual Opteron 2427 2.2 GHz (12 cores total), 8 GB RAM, Centos Linux.
pachi, running on 64 platforms, each x86 64 bits, 32 GB ram, using 22 cores of each, giving about 1500 playouts/s/core at the beginning of a 19x19 game.
Fuego, running on an 8-core i7 920 (like CZechBot)
running on one processor of a 4GiB RAM, AMD Athlon(tm) 64 X2 Dual Core Processor 4000+, shared with WeakBot50k
Steenvreter, running on an Intel Core2, Q6600 (4 cores)
Zen, running on a Mac Pro 8 core, Xeon 2.26GHz