Eighty-fourth KGS Computer Go Tournament

Sunday July 1st, 2012

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

Crazy Zen19 pachi AyaMC ManyF nomit Fuego gomor Orego MCark
B14 W07R W12R W11R W13R W16R W15R 63024Winner
2Zen19S W04 B17R
B15R W12R B16R W13R W11R 62822
3pachi B02R W05R
W17R B14R W13R B11R B16R 52513
4AyaMC B01R B02R B07R
W14R B15R W13R B16R 42710
5ManyFaces2 B03R W06R W04R
B12R W11R W17R B15R 42710
6nomitan440 B06R B03R B04R W02R
W17R B11R B15R 3255
7Fuego19 B03R W05R B01R B07R
B16R B12R B14R 3204
8gomorra32 B05R B03R B07R W01R W06R
B14R B12R 2222
9Orego12 W01R W06R W05R W02R W04R
B13R W17R 2180
10MCark B01R W06R W05R W04R W02R W03R B07R

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.

Crazy Stone

The organiser and players were pleased to see Rémi Coulom enter his program Crazy Stone for this event. Crazy Stone had not played in a KGS event since April 2010, but was known to be one of the strongest players, similar in strength to the impressively strong Zen.

The hazards of the first day of the month

The first day of a month is an unfortunate date for these events. Three months ago, on April 1st, KGS's celebration of April Fool's Day make things difficult for me, as I described in my April report. On July 1st, or rather, between the end of June 30th and the start of July 1st, UTC, there was a leap-second, which confused many unix systems around the world. One such system was the KGS server, which started to suffer from serious lag. At the time, the TCGA 19x19 Computer Go Tournament was taking place, and the lag caused all its games from then on to be lost on time, or by forfeit.

As it happened, this was convenient for me. On the evening of June 30th, I was aware of a problem with the scheduling of this tournament. The TCGA Tournament should have been over before this July KGS tournament was scheduled to start. However, the KGS scheduler, instead of scheduling four games each round to the eight players, was scheduling three games and two byes; and as it was a round robin, this was going to make it overrun, and end at least an hour after the KGS July tournament was due to start, with many of the same players. When I went to bed on the night of June 30th, I did not know how to resolve this clash. But when I woke up on the morning of July 1st, I found that the TCGA tournament was over, with all the games of its final 28 rounds ending as time losses or by forfeit. While this was most unfortunate for the TCGA organisers and players, it resolved my problem.

When the KGS July tournament started, the cause of the lag on KGS was still unknown, and there were fears that it might recur. In fact, wms' action in rebooting the server had fixed the lag, but as its cause was still not known, some were worried that it might recur.

Once we realised that the lag had been due to the leap-second, we realised that it would not recur. In fact, as pasky suggested, it could have been fixed without a server reboot, by the unix command

  date -s "`date`"
  date -s $(date)


I have nothing to say about the play in this event. Most of the players are much stronger than me, and I noticed no poor moves.

Orego12 was missing when round 1 started, and arrived 11 minutes late for its game with pachi, which it lost.

After round 3, CrazyStone and Zen19S were tied for first place, with three wins each. In round 4 they played each other, and CrazyStone won by the minimal margin of 1½ points. Remarkably, this was the only game in the whole tournament not won by resignation.

In round 7 Zen19S and CrazyStone played again. This time Zen won.

Annual points

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

Crazy Stone7
Many Faces of Go

Details of processor numbers, power, etc.

Aya, running on 6 cores of an i980X, at 3.3GHz.
Crazy Stone, running on 24x2.2GHz AMD Opteron.
Fuego, running on two Intel Xeon X5670, 12 cores (24 threads) at 2.93 GHz.
Gomorra, running on an Infiniband cluster using 32x12cores. Cores are running at 2.67 Ghz.
Many Faces of Go, running an Amazon EC2 cc2.8xlarge instance (ES2670 CPU) using 8 cores, 8 threads, 24 GB memory.
MC_ark, running on a Core-i7 2600K 3.40GHz*4core (8 threads)
nomitan, running on a cluster of 18 servers, with a total of 160 cores.
Orego, runing on one node of a custom Linux cluster. The node has two AMD Six Core Dual Opteron 2427 2.2 GHz (12 cores total), 8 GB RAM, Centos Linux.
pachi, running as 74 threads over 11 machines.
Zen, running on a mini computer cluster consisting of a dual 6-core Xeon X5680@4.2 GHz 24 GB RAM, a 6-core i7 3930K@4.2 GHz 16 GB RAM, a 6-core Xeon W3680@4 GHz12 GB RAM, and a 6-core i7 980X@4 GHz 6 GB RAM PCs connected via a GbELAN. 4 PCs (30 cores) total.