Fifty-sixth KGS Computer Go Tournament

Sunday February 7th 2010

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

Rules

format10-round Swiss
board size19×19
rulesChinese
komi
time14 minutes plus 25/20s

Times

The first round started at 08:00 UTC.

Result table


-
placenamewinsSOSSoDOS
1stZen19105454
2ndManyFaces175629
3rdstv75130
4thAyaMC65725
5thPueGo64717
6thpachi255822
7thOrego54510
8thPNUGo3504
9thvalkyria192512
10thbreak192410

The numbers in this table may not add up as you
might expect. This is because byes, and games
against absent opponents, are both treated as wins.

We welcomed a new person to this event: Jean-loup Gailly, who is working on pachi. His contributions to the pachi2 that competed included improvements to time management and to memory management, but did not directly involve the move generating code.

Results

In round 1, stv's operator stated "kgsgtp disconnected stv without a clear reason". He was able to reboot stv, and it resumed and won its game against break19, despite some lost time.

In round 2, valkyria19 was disconnected by something local to it. It was rebooted, and resumed its game against AyaMC, but was now short of time, and lost on time.

CzechBot (a build of MoGo) had registered for the event, did not appeared on KGS. Towards the end on round 3, I decided that it was not going to appear, and removed it from the draw.

In round 4, pachi2 timed out in its game against PueGo, in a lost position. It had taken between four and six seconds for each of its previous ten moves, then used all its remaining 211 seconds without moving at all, so it seems likely that something was wrong.

PueGo vs AyaMC
After move 277.

In the round 8 game between PueGo and AyaMC SGF they reached the position shown to the left. White, PueGo, is a few points ahead here even if it passes. The black group at the top left is dead (it has only one eye), and the black group midway down the left side is very dead (it has only one liberty). However, PueGo played at 1 filling a vital own liberty, and continued with 3, which loses the game. On move 4, Puego's estimate of its probability of winning dropped from 0.92 to 0.07 and it resigned.

Zen19 vs AyaMC
Moves 65-85.

In the round 10 game between Zen19 and AyaMC SGF AyaMC as Black played out a losing ladder, as shown to the right. This gave White a huge advantage, which in typical MC fasion it converted into a half-point win. (Technically, AyaMC lost on time on move 454, 167 moves after the last dame had been filled.)

Processor numbers, power, etc.

AyaMC
Aya, running on Xeon 2.6GHz 8cores
break19
break, probably running on a single processor Intel(R) Celeron(R), 1.7Ghz
ManyFaces1
Many Faces of Go, running on a quad core Q8200, 2.3 GHz.
Orego
Orego, unspecified but probably running on 2 x 3 GHz Dual-Core Intel Xeon
pachi2
pachi, probably running on 23 cores (2.4 GHz) until round 4, and then 13 cores.
Puego
Fuego r1096, running with four threads on an i7 920.
PNUGo
GNU Go, unspecified platform
stv
Steenvreter, running on an Intel Core2, Q6600
valkyria19
valkyria, running on a single processor Pentium M 1.4Ghz
Zen19
running on a Mac Pro 8 core, Xeon 2.26GHz.