Past computer-computer results | future events | computer-go.info

Human-Computer Go Challenges

This page lists Go matches between humans and Go-playing programs.

The programs appearing in this list are generally the strongest, or among the strongest, of their time. In many cases the program is one which has just won a competition among programs.

The humans are mostly strong players, many of professional or near-professional rank. Their ranks, where known, are listed. Where results are given, the human's wins are given first, so "3-1" means that the human won three games, the computer won one.

Thus this page gives some indication of the increasing strength of Go programs.

Graphs

The data in the table below is summarised in graphical form here.

Warning

Caution is needed in drawing conclusions from the data presented here. In particular, it is possible for a human to observe, and learn to take advantage of, the weaknesses in the play of a program. Someone who persistently plays the same program can improve his results against it greatly; by nine stones or even more. For example, in the table below we see that in 1995, two of three inseis (about amateur 6d) who had played HandTalk only once before, lost to it giving 13-stone handicaps; but in 1998 Jean-loup Gailly, then amateur 5-kyu, was able to give it 20 stones and beat it. He was playing the commercially-available version, which may have been weaker than the version that competed for the Ing prize; but it is probably more significant that he had spent many games investigating and learning to exploit its weaknesses.

This warning has been rather less relevant from 2008, with all leading programs using Monte Carlo techniques. It is more difficult, though still possible, to learn to take advantage of the weaknesses of a Monte-Carlo program. The play of these programs has a significant random element, so you can never be sure that "in this position it will make this blunder" and arrange to set up such a position.

List of games, latest at top

1986 | 1987 | 1988 | 1989 | 1990 | 1991 | 1992 | 1993 | 1994 | 1995 | 1996 | 1997 | 1998 | 1999 | 2000 | 2001 | 2002 | 2003 | 2004 | 2005 | 2006 | 2007 | 2008 | 2009 | 2010 | 2011 | 2012 | 2013 | 2014 | 2015 | 2016

Date/linkHumanProgramsizehandicapResultEventcomments
       2018
2018-07-27 – 08-12 undecided undecided 19×19 undecided European Go Congress, Pisa, Italy The organisers plan to include "Computer Go (Exhibition & Competition)".
       2017
2017-08-18 "A top human professional" The winner of the 2017 CITIC Securities Cup 19×19 even Ordos City, Inner Mongolia, China Sponsored by CITIC Securities
2017-08-02 Ohashi Hirofumi 6p
Tang Yi 3p
Zen19X
LeelaX
19×19 even 1-1 SGF SGF EGC, Oberhof, Germany The games were broadcast on KGS using usernames DGOBn, starting at 13:00 UTC.
2017-07-12 Chou, Chun-Hsun 9p Zen
CGI
19×19 even 0-2 SGF SGF
0-2 SGF SGF
FUZZ-IEEE 2017, Naples, Italy There were also three "human prediction games" involving DarkForest and Chun-Hsun Chou 9p.
2017-06-19, -21 Shin, Minjun 5p (Korea)
Wang, Haoyang 6p (China)
DeepZen 19×19 even 1-1 SGF SGF Third "Dream Mlily Cup" World Open Go Championship This is a world Go championship. In previous years, only humans have been able to enter, but in 2017 DeepZen was given a wildcard entry into the top 16 places.
2017-05-23 – 27 Ke Jie, various Chinese professionals AlphaGo 19×19 even 0-3 SGF SGF SGF "Future of Go Summit", Wuzhen, China Three events:
  1. Ke Jie vs AlphaGo, a three-game match
  2. "Pair Go", each side being a professional and AlphaGo
  3. "Team Go", AlphaGo vs a team of five professionals
2017-03-26 Ichiriki Ryo 7p The winners of the 2017 UEC Cup: JueYi (Fine Art) and DeepZenGo 19×19 even 0-2 SGF SGF 5th Densei-Sen
2017-03-21, 22, 23 Top three professional players from China (Mi Yu Ting 9p), Korea (Park Jeong Hwan 9p) and Japan (Yuta Iyama 9p). DeepZenGo 19×19 even 2-1 SGF SGF SGF World Go Championship, organised by the Nihon Ki-in

Time limits three hours each, including five one-minute byo-yomi periods.

This was not primarily a human/computer event. It was a world championship, which in 2017 had three human contenders and one computer.

       2016
2016-11-23 Yongil Ha 6p Zen 19×19 even 0-1 SGF Neyagawa Igo Shogi Festival Played on KGS. Time setting was 30 second a move, no main time.

Zen ran on a dual Intel Xeon E5-2687W v3 3.1 GHz server with an nVidia GTX-1080. Only seven threads ran, because of one GPU.

2016-11-19, -20, -23.

Cho, Chikun 9p

With commentaries by
Game 1: Ichiriki, Ryo
Game 2: Takao, Shinji
Game 3: Iyama, Yuta
and, in English, Michael Redmond and Antti Tormanen.

Zen 19×19 even 2-1 SGF SGF SGF Second Go Denou-sen

Broadcast on Niconico:
Game 1 (EN),
Game 2 (EN),
Game 3 (EN).

Three-game series. Time limits were two hours each plus three 60-second byo-yomi periods.

Zen used CPU: 2 x Intel Xeon E5-2699v4 (44 cores/2.2 GHz), GPU: 4 x nVidia Titan X (Pascal), RAM: 128GB.

2016-08-28, -09-04, -09-11, -09-18. Lukas Krämer 6d Zen (the winner of July's ICGA tournament) 19×19 even 1-3 SGF SGF SGF SGF Codecentric Challenge Five-game match, ended when one player had three wins. Each game was played on KGS.
2016-08-23 Chou, Chun-Hsun 9p Zen 19×19 Two stones 1-0 SGF ICIRA

Zen was running on a dual-xeon server with no GPU installed. It was not using a value network.

2016-08-04 Andy Liu 1p DarkForest 19×19 Three stones 1-0 SGF 2016 US Go Congress
2016-07-27, starting at 16:00 UTC Cho Hye-Yeon 9p Zen 19×19 Two stones 0-1 SGF 2016 European Go Congress
2016-07-24, 20:00 UTC Chou, Chun-Hsun 9p DarkForest 19×19 2 stones 1-0 SGF IEEE WCCI 2016

Played in Vancouver, Canada.

Some of the games are available on KGS, in the accounts 'wcci201601' and 'ZensGuest'.

Ping-Chiang Chou 6p CGI 2 stones 1-0 SGF
Sheng-Su Chang 6d Zen 2 stones to human 0-1 SGF
22:00 UTC Chou, Chun-Hsun 9p CGI 2 stones 1-0 SGF
Ping-Chiang Chou 6p DarkForest 2 stones 1-0 SGF
Shang-Rong Tsai 6d pachi 2 stones 1-0 SGF
Shi-Jim Yen 6d Zen 2 stones to human 0-1 SGF
2016-07-25, 16:30 UTC Ping-Chiang Chou 6p Yi-Min Hsieh 6p and CGI even 1-0 SGF
Chou, Chun-Hsun 9p Zen 2 stones 1-0 SGF
Shi-Jim Yen 6d DarkForest 2 stones to human 0-1 SGF
19:30 UTC Chou, Chun-Hsun 9p Zen 3 stones 0-1 SGF
Shang-Rong Tsai 6d CGI 2 stones 1-0 SGF
Sheng-Su Chang 6d DarkForest 2 stones to human 0-1 SGF
21:30 UTC Shang-Rong Tsai 6d DarkForest even cancelled
Sheng-Su Chang 6d CGI 2 stones to human 1-0 SGF
volunteer, 6d+ Zen 2 stones to human 0-1 SGF
23:00 UTC Chou, Chun-Hsun 9p Yi-Min Hsieh 6p and DarkForest even 1-0 SGF
Ping-Chiang Chou 6p CGI 2 stones 0-1 SGF
Shang-Rong Tsai 6d Zen even cancelled
2016-07-26, 15:30 UTC Chou, Chun-Hsun 9p Yi-Min Hsieh 6p and DarkForest even 1-0 SGF
Ping-Chiang Chou 6p Zen 2 stones 0-1 SGF
Shang-Rong Tsai 6d CGI no komi 0-1 SGF
Sheng-Su Chang 6d CGI 2 stones 1-0 SGF
2016-06-07 Yoko Takemiya 6p Zen 19×19 Two stones 0-1 SGF 30th Annual Conference of The Japanese Society for Artificial Intelligence Played in Kokura, Japan. The time-setting was 60 minutes plus 30 seconds byoyomi. Japanese rules were used.
2016-06-05 Hajin Lee 3p Crazy Stone 19×19 even 1-0 SGF The game was played on KGS.
2016-04-28 Aoi Kumamoto Zen 19×19 Even 1-0 SGF "Good Life Festa", Chukyo TV.
See also Kumamoto family Go blog.
Played in Nagoya, Japan. The time-setting was (no main time and) 30 seconds byoyomi. Japanese rules were used.
2016-03-23 Kobayashi, Koichi 9p Zen 19×19 3 stones 0-1 SGF 4th Densei-Sen, in Tokyo These were the two highest-placed players in the 9th UEC Cup.
DarkForest 1-0 SGF
2016-03-19 Hisahiro Yoshizaki (8d?) Zen (the highest-placed program in the 9th UEC Cup) 19×19 Even 0-1 9th UEC Cup (JP), in Tokyo
2016-03-08, -09, -11, -12, -14 Lee, Sedol 9p,
world champion
AlphaGo 19×19 Even 1-4
SGF SGF SGF SGF SGF
DeepMind AlphaGo vs Lee Sedol

이세돌 9단-구글 인공지능, 중국 바둑규칙으로 대국
Held in South Korea, games start at 22:00 UTC.
Chinese rules, 7½ komi.
2 hours each plus 3 1-minute byoyomi periods.

Games were broadcast live on KGS.

Commentary by Li, Zhe (6p)
2016-02-07 Atsushi Ida 8p Zen 19×19 4 stones 0-1
SGF
Sponsored by the Nihon Ki-in (JP) Zen ran on a dual Xeon E5-2687W v3 server, 2×10 3.1 GHz cores, as used by Zen19X on KGS.

Time limits were 30 minutes + 30 seconds byoyomi.
       2015
2015-11-14 Lian, Xiao 7p DolBaram 19×19 4 stones, 5 stones, 6 stones. 2-1
SGF SGF SGF
A three-game match. After each game, the handicap changed by one stone, up or down depending on the result of the previous game.
2015-10-03, -17, -24, -31 Franz-Josef Dickhut 6d (GoR=2547) Zen 19×19 Even 3-1
SGF SGF SGF SGF
codecentric Go Challenge, 2015 Best-of-five match, played on KGS, and sponsored by codecentric AG

Time limits 60 minutes plus 40 seconds a move byo-yomi.
2015-10-10 Pocsai, Rita 5d (GoR=2429) Zen 19×19 Even 2-0 SGF SGF DACHS Go Challenge In Budapest. Sponsored by DACHS
Players had an hour each plus generous byo-yomi.
2015-10-11 Rigó, István 3d (GoR=2206) 0-1 SGF
Played 2015-10-05 – 09.

Announced 2016-01-27
Fan, Hui 2p (GoR=2759) AlphaGo 19×19 Even. 0-5
SGF SGF SGF SGF SGF
As reported in Nature AlphaGo was created by a team within Google DeepMind. It used 1202 cpus and 48 gpus.

Time limits were one hour each main time, one move in 30 seconds overtime.

See also: Nature article, Google Research Blog, BGA News, USGO News, a Youtube video; game 1, game 2, game 3, game 4, game 5.
2015-08-31 Chou, Chun-Hsun 9p,
Chang, Kai-Hsin 5p,
Yu, Li-Chun 1p
Zen 19×19 4 stones 1-0, 0-1, 1-0 IEEE CIG 2015 Each professional played each program once.

The time-settings were all 45 minutes each.
Aya 5 stones 0-1, 1-0, 0-1
CGI 6 stones 0-1, 1-0, 1-0
Yung-Hao Hung (6D) Zen even 1-0
Bo-Yuan Hsiao (6D) Aya no komi 0-1
Chi Chang (5D) CGI 1-0
2015-08-02 Wen-Chih Chen 6d Zen 19×19 None 1-2 SGF SGF SGF FUZZ-IEEE 2015 Played in Istanbul, Turkey.
The time-setting was 45 minutes each.
2015-08-03 Ping-Chiang Chou 6p 4 stones 2-0 SGF SGF
5 stones 0-1 SGF
Shi-Jim Yen 6d CGI 4 stones 0-2 SGF SGF
2015-08-04 Ching-Nung Lin 6d Jimmy 4 stones 2-0 SGF SGF
5 stones 1-0 SGF
Francesco Marigo 4d CGI 2 stones 2-0 SGF SGF
3 stones 1-0 SGF
2015-07-29 Hajin Lee 3p AyaMC 19×19 Five stones 0-1 SGF 2015 European Go Congress Played in Liberec, Czech Republic.
The time-setting was 60 minutes each plus byoyomi. Japanese rules were used.
As Hajin Lee played the game, she gave a live commentary on it, published on YouTube.
2015-05-31 Miori Shimosaka 2p Zen 19×19 Three stones 1-0 SGF 29th Annual Conference of The Japanese Society for Artificial Intelligence JP Played in Hakodate, Japan.
The time-setting was 60 minutes plus 30 seconds byoyomi. Japanese rules were used.
2015-03-17 Chikun Cho 9p Crazy Stone 19×19 3 stones 1-0 SGF Densei-sen, in Japan. video, video These were the two highest-placed players in the 8th UEC Cup.
DolBaram 4 stones 0-1 SGF
2015-03-15 Masato Sakakibara (7d, Tygem) Crazy Stone 19×19 even 1-0 SGF The Exhibition Match of the 8th UEC Cup JP Japanese rules were used, with 6½ komi and 40 minutes each sudden death.
       2014
2014-11-07 Meien O 9p Aya 9×9 even ½-½ SGF GPW Cup, in Japan

Apart from fuego, these were the players in the Game Programming Workshop (GPW). (Storm and Zen had server problems and could not play.)

The programs all took black, and gave seven points komi. O Meien's time was unlimited. All programs ran on laptop computers, typically 2 GHz dual core.

O Meien might have beaten Aya if he had been right in his belief that Japanese rules were in use.

Coldmilk1-0 SGF
Fuego1-0 SGF
MC_ark1-0 SGF
Nomitan0-1 SGF
Ray1-0 SGF
Best-of-five match: 2014-10-04, -11, -18, -25 (-11-08) Franz-Josef Dickhut 6d (GoR=2537) Crazy Stone 19×19 Even 3-1 SGF SGF SGF SGF codecentric go challenge 2014
The challenge
Interview
Played on KGS.

Time limits one hour each plus byo-yomi

2014-07-26 Kaoru Yosano Zen 19×19 even 0-1 SGF Niconico Igo Circle, final special
2014-05-13 Kenta Sakakura (est. 8d) Zen 19×19 Two stones; komi of 6½ points to Black 0-1 SGF 28th Annual Conference of The Japanese Society for Artificial Intelligence JP Played in Matsuyama, Japan.
The time-setting was 60 minutes plus 30 seconds byoyomi. Japanese rules were used.
2014-03-21 Norimoto Yoda 9p Crazy Stone 19×19 4 stones 0-1 SGF 2nd Densei-sen, in Japan
Zen1-0 SGF
2014-03-16 Yoshikuni Tanaka (est. 6d) Zen 19×19 even 0-1 SGF 7th UEC Cup, in Japan. These were the two highest-placed players in the 7th UEC Cup.
Each player had 30 minutes.
2014-02-11 Li Yu Chang, 8p Zen 9×9 even 2-0 SGF SGF Go Denosen 20 minutes each, 30 seconds byo-yomi
Tomoya Hirata, 3p 2-0 SGF SGF
2014-02-16 Kiko Emura, 7d 13×13 2-0 SGF SGF 30 minutes each, 30 seconds byo-yomi
Ichiro Ozawa 19×19 0-1 SGF 60 minutes each, 60 seconds byo-yomi
       2013
2013-07-09 Ping-Chiang Chou 5p MoGoTW
Many Faces of Go
Zen
Coldmilk
19×19 4 stones,
5 stones
2-0 SGF SGF FuzzIEEE2013

in Hyderabad, India

Games were played on KGS, starting at 05:30 UTC.
9×9 even 4-3 SGF SGF SGF SGF SGF SGF SGF
Shi-Jim Yen 6d 19×19 2 stones,
3½ reverse komi
1-1 SGF SGF
19×19 even 5-5 SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF
Ching-Nung Lin 6d 19×19 2 stones,
no komi
1-1 SGF SGF
19×19 even 3-4 SGF SGF SGF SGF SGF SGF SGF
2013-06-05 Takuto Ooomote (9d on Tygem) Zen 19×19 3 stones 0-1 SGF 27th Annual Conference of The Japanese Society for Artificial Intelligence
JP EN
Played in Toyama, Japan.

The time-setting was 60 minutes plus 30 seconds byoyomi. Japanese rules were used.

2013-03-20 Ishida, Yoshio 9p Crazy Stone, the winner,
and Zen, the runner-up,
of the UEC Cup
19×19 4 stones 1-1
SGF
SGF
Denseisen Den-sei-sen (literally "electronic holy wars") will run for five years, starting in 2013. It is a co-operation between the UEC and the Nihon Ki-in. The winners of the UEC Cup will play exhibition games against professional players.
2013-03-17 Bungo Taga (estimated 9d) Crazy Stone 19×19 even 1-0 SGF The 6th UEC Cup Each player had 30 minutes.
       2012
2012-11-26 ? MoGoTW ? ? The games may have been broadcast on KGS
2012-11-25 So, Yokoku (a.k.a. Yaoguo Su, Y. Soo) 8p

Ohashi, Hirofumi 5p

Ichiriki, Ryo 2p
Zen 9×9 even 2-0 SGF SGF

2-0 SGF SGF

2-0 SGF SGF
(in Japanese) The time setting was 20 minutes main time and 30 seconds per move (no extra byo-yomi). Chinese rules and 7.0 komi were used. Zen's hardware was 4 pc cluster (12, 6, 6 and 6 cores) at 4 GHz, as for TAAI and Zen19D/S on KGS.
2012-11-23 Yongil Ha 5p Zen 19×19 4 stones 0-1 SGF Neyagawa Igo Shogi Festival Japanese rules with ½-point komi. Time limit was 30 seconds a move. Zen used a home-made 30-core network.
2012-11-17 Chou, Chun-Hsun 9p

Chang, Kai-Hsin 4p

Zhao, Pei-Zhe 3p
Zen

Many Faces of Go

AyaMC
19×19 4 stones
4 stones, 5 stones
4 stones, 5 stones
0-2 SGF SGF

2-0 SGF SGF

1-1 SGF SGF
TAAI 2012, Tainan, Taiwan

The games were broadcast on KGS. The professionals used accounts taai1, taai2, taai3, taai4.


Chou, Ping-Chiang 5p Many Faces of Go 13×13 ½ komi, 3½ komi to B 2-0 SGF SGF
Coldmilk 9×9 Even, 7 komi 1-1 SGF SGF
2012-08-01 Catalin Taranu 5p Crazy Stone 19×19 4 stones 1-1 SGF SGF European Go Congress, Bonn, Germany
Motoki Noguchi 7d Zen 13×13 even 1-1 SGF SGF
2012-07-31 9×9 even 0-2 SGF SGF
2012-06-13 Chou, Chun-Hsun 9p Zen 19×19 4 stones 0-1 SGF IEEE WCCI 2012, Taipei, Taiwan. All the games were played on KGS.
an amateur 6d 13×13 even 1-0 SGF
9×9 even 0-1 SGF
2012-06-12 Joanne Missingham 6p 19×19 4 stones 1-0 SGF
2012-03-17 Takemiya, Masaki 9p Zen 19×19 5 stones 0-1 SGF 6th E&C Symposium Zen used a mini-cluster of four PCs, with a total of 22 cores.

The games were be broadcast on Niconico.

Time limits were 3 hours each for 19×19, 20 minutes each for 9×9, both with 30 seconds byo-yomi.

4 stones 0-1 SGF
Ohashi, Hirofumi 5p 9×9 even (7 points komi) 1-1 SGF SGF
2012-03-06 Motoki Noguchi 7d
Hosoya, Takashi 7d
pachi2
Zen
13×13 even 1-0 SGF
0-1 SGF
2012 JAIST Cup
2012-02-25 Ishida, Yoshio 9p Zen 13×13 komi only 0.5 1-0
SGF
Time limits 25 minutes + byo-yomi. Ishida said Black wasted ko threats before ko fight.
2012-01-14 to 2012-01-17 John Tromp 2d Zen 19×19 even 1-3
SGF SGF SGF SGF

A rerun of the 2010 Cook/Tromp $1000 bet.

5-game match, played on KGS. Time limits 90 minutes each + byo-yomi.

Zen used a dual 6-core Xeon X5680, over-clocked to 4.2Ghz.
       2011
2011-12-03, 04 Chizu Kobayashi 5p Erica 19×19 6 stones 1-0 SGF 5th UEC Cup, Tokyo
Meiko Tei 9p Zen 0-1 SGF
2011-08-03 Kozo Hayashi 6p Zen 19×19 5 stones 0-1 SGF European Go Congress, Bordeaux, France Zen used a cluster of six computers with a total of 26 cores.
Motoki Noguchi 7d 13×13 ½ point komi 0-1 SGF
9×9 even 2-0 SGF SGF
2011-06-27, 28 Chou, Chun-Hsun 9p
Chou, Ping-Chiang 5p
Missingham, Joanne 5p
Chang, Kai-Hsin 4p
Fuego
MoGoTW
Many Faces of Go
Zen
19×19 6 stones 6-2 SGF FUZZ-IEEE 2011, Taipei, Taiwan.

All the games were played on KGS.

Fuego used 12 cores.
Many Faces of Go used 12 cores.
MoGoTW used 720, 712 and 48 cores for different events.
Zen used 26 cores.
19×19 rengo 3-1
13×13 two stones, 3½ komi 4-4
9×9 even (7½ komi) 6-2
Ten professionals, rated from 1p to 5p MoGoTW 9×9 even (7½ komi) 21-9 MoGoTW used 64 cores.
7×7 even (9½ komi) 0-20
2011-06-11 Kim Young Sam 8p MyGoFriend 9×9 even 0-2 SGF SGF

"Showmatch" played on KGS.

This was scheduled as a two-game match. After MyGoFriend won two games, they played two more games, won by Kim.
2011-05-04 Chou, Chun-Hsun 9p
Chou, Ping-Chian 5p
MoGoTW 9×9 even 1-1 SGF SGF
0-2 SGF SGF

Human vs. Computer Go Competition, Paris, France.

The bots were running on 40-core clusters.

These 9×9 games were "blind go" – as in blindfold chess, the human players did not see a board, they had to remember the position.
Chou, Chun-Hsun 9p
Chou, Ping-Chian 5p
Chou, Ping-Chian 5p
Chou, Chun-Hsun 9p
MoGo 13×13 2 stones
2 stones
2½ stones
2½ stones
1-0 SGF
1-0 SGF
1-0 SGF
0-1 SGF
"2½ stones" means three stones but 3½ points of reverse komi.
Chou, Chun-Hsun 9p and Chou, Ping-Chian 5p (rengo)
Chou, Chun-Hsun 9p
Chou, Chun-Hsun 9p
Chou, Ping-Chian 5p
Chou, Chun-Hsun 9p

MoGoTW and Pachi
MoGoTW
Pachi
Pachi
MoGoTW
19×19
6 stones
7 stones
7 stones
6 stones
6 stones

1-0 SGF
1-0 SGF
0-1 SGF
1-0 SGF
1-0 SGF
2011-03-22 Ohashi 4p Aya, Erica, Fuego, MyGoFriend 9×9 even Plans for this event were abandoned following the Sendai earthquake.
2011-03-09 Kai-Shin Chang 3p Yin-Nan Chou 3p Meng-Yun Yang 3p Yu-Hsiang Lin 3p Chun-Yen Lin 2p Pei-Je Chao 2p Shao-Jie Ting 2p Cheng-Jui Yu 1p Yu-Bang Kou 1p Yen-Chen Huang 1p MoGoTW 9×9 The bots always played White, with komi sometimes 7½, sometimes 8½ 14-7
SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF
Played in Taiwan, starting at 10:30 local time, = 02:30 GMT, using KGS.

Time limits were 30 minutes each, absolute.

Each pro played two games with MoGoTW (one pro played three)
2011-02-15 Chun-Shung Chou 9p
Ping-Chiang Chou 5p
Cheng-Rui Yu 1p
MoGoTW 9×9 even 7-1
SGF SGF SGF SGF SGF SGF SGF SGF
Played in the University of Tainan, Taiwan, starting at 04:20 GMT. All 30m.
'mogotw1' had 24 cores and 64GB of RAM, 'mogotw2' 16 cores and 36GB, and 'mogobot5' 16 cores and 32GB.
       2010
2010, 28-29 December John Tromp 2d Many Faces of Go 19×19 even 4-0
SGF SGF SGF SGF
7-game match Played to settle a bet.
2010-11-28 Kaori Aoba 4p Zen 19×19 6 stones 0-1 4th UEC Cup in Tokyo. Zen was running on a PC cluster (6 x 4 GHz Intel Xeon and 5 x 4 x 3 GHz Intel Core2).
Meiko Tei 9p Fuego 1-0  
2010-10-01 Rina Fujisawa 1p Erica 19×19 6 stones 1-0 Played at the Computer Olympiad in Kanazawa, Japan. Fujisawa (aged 12) reported "Computer played non-book move in middlegame, and blew up."
2010-07-28 Catalin Taranu 5p MoGoTW 9×9 Even (7½ komi) 1-0 European Go Congress in Tampere, Finland. MoGoTW was running on 512 cores of the Louhi Cray XT4/XT5 supercomputer.
19×19 Seven stones 0-1
2010, 18-23 July Chun-Hsun Chou 9p
Ping-Chiang Chou 4p
Shang-Rong Tsai 6d
Shi-Jim Yen 6d
Fuego, 110 cores

Many Faces of Go, 12 cores

MoGo/MoGoTW, 120 cores

Zen

19×19 7 stones for 9p 2-0 SGF SGF A series of games with pros and strong amateurs, at the WCCI in Barcelona, Spain. 45m
6 stones for 4p 1-1 SGF SGF 45m. Zen beat Ping-Chiang Chou 4p
13×13 2 stones 3-1 SGF SGF SGF SGF 45m. MoGo 0-2, MF 1-1, Fuego 2-0.
9×9 even 7-5 SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF SGF 30m. The pros won all their games as white and some as black. Mogo scored 1-1 vs Chou 9p and 0-2 vs Chou 4p. Fuego scored 0-2 vs Chou 9p and 1-1 vs Chou 4p. Zen scored 3-1 vs the 6ds.
2010-04 Alexander Lücking 2d Many Faces of Go 13×13 Alexander played white 1-0 SGF Mensch gegen Maschine Schaupartie, Paderborn
2010-02-06 Nam Chi-hyeuong 1p Zen 9×9 even 0-1 SGF Played on KGS for a symposium at Osaka Syogyo University. No main time, 30 seconds byo-yomi. Zen was running on an 8-core Mac Pro, 2.26 GHz.
       2009
2009-12-31 'mosa' 5d Zengg 19×19 even 3-1 SGF Not an organised event – four games played on KGS. 'mosa' has a well-established 5d KGS rating. Played with 30m+5*30s. Mosa won the other three games, only the one he lost is shown here.

Zengg was running on a 16-core cluster.

2009-12-06,07,08 Li Yue 6d MoGo 19×19 6 stones 1-0 SGF 'Man vs MoGo' challenge at Cádiz Open Go Tournament. 19×19; played with no main time, 45 seconds byo-yomi

MoGoTW was running on 2 Power6 nodes, with 32 cores per node.

9×9 played with 50 minutes each.

Chun Poong Jho 8p 1-0 SGF
9×9 even 1-1 SGF SGF
2009-11-29 Kaori Aoba 4p Zen 19×19 6 stones 1-0 SGF Played after the 3rd UEC Cup (UEC Cup, English version) All the audience believed Zen would win, but Zen lost a ko near the end, and with it the game.
Cheng Ming Huang 9p (Japanese name: Meiko Tei) KCC Igo 1-0 SGF White won the game with no trouble.
2009-10-26 Chun-Hsun Chou 9p MoGoTW 9×9 even 2-1 SGF SGF SGF Played during a press conference at Taipei The game which MoGoTW won was played with 45 minutes each. The other two were played with 10 minutes each and with 15 minutes each.

MoGoTW was running on 32 quad-cores.

2009-10-24 Javier-Aleksi Savolainen 5d MoGo version 4.86 9×9 even 0-1 SGF At the Alternative Party digital culture festival in Helsinki, Finland. MoGo was running on a 56-core Cray CX1 available locally at the festival.
19×19 3 stones 1-0 SGF
2009-08-29 Fan Hui 2p (GoR=2759) MoGo 9×9 even 2-1 SGF SGF SGF At the French championship, in Toulouse. MoGo was running on a 16-core R90, at 3GHz, with 32Gb of RAM.

The second game record is incomplete, or MoGo resigned.

The third game was the same as the second until move 13, when Fan Hui (again playing black) tried to improve on his play in the previous game; and lost.

2009-08-21, 2009-08-22 Chun-Hsun Chou 9p Many Faces of Go 19×19 7 stones 1-0 SGF Played on Jeju Island ("Quelpart"), South Korea, at the 2009 IEEE International Conference on Fuzzy Systems. Fuego running on 80 cores: ten 8-core nodes each with two quad core Xeon E5462 2.8GHz processors and 32Gb of main store. A 20Gbps network connects the nodes.

Many Faces of Go running on 32 cores: 4 nodes each with 2 quad core Intel Xeon (x5460) running at 3.16 GHz with 16Gb of RAM, a total of 64 Gb.

MoGo running on 640 cores: supercomputer "Huygens" with 20 nodes each of 32 cores.

Zen running on 8 cores: Mac Pro with 8 core processors (Quad-Core Intel Xeon 2.26GHz x2).

Zen 1-0 SGF
MoGo (extra game) 1-0 SGF
Shen-Su Chang 6d MoGo 4 stones 1-0 SGF
Fuego 0-1 SGF
Chun-Hsun Chou MoGo 9×9

Two-game matches.

even 1-0 SGF
Fuego 1-1 SGF SGF
Shen-Su Chang 6d Many Faces of Go 2-0 SGF SGF
Zen 0-2 SGF SGF
2009-08-10 O Meien 9p Zen 9×9 Komi of 3½, 2½, 3½ 2-1 SGF SGF SGF Played in Tokyo, broadcast on KGS using account 'oumeien'. Zen was running on a Mac Pro, with eight Xeon cores. For the 19×19 it was set to use 10 seconds per move; for the 9×9, 15 seconds. Zen lost the 9×9 when the komi was 2½, and won when it was 3½.
19×19 7 stones 1-0 SGF
2009-08-07 Myungwan Kim 8p Many Faces of Go 19×19 7 stones 1-0 (SGF) SGF Played at the US Go Congress. Each player had 45 minutes. Many Faces of Go was running on a 32-core system. The first, incomplete, game was a test game.
2009-05-22 Catalin Taranu 5p MoGo 9×9 even 3-1 SGF SGF SGF SGF At Espace des Sciences in Rennes, France In its first three games, MoGo consumed a lot of time on its early moves. After it lost these games, its operators changed its time settings to make it keep more time for the middle game.
2009-02-14 James Kerwin 1p Many Faces of Go 19×19 7 stones 0-1 SGF 2009 AAAS Annual Meeting Many Faces of Go was running on a 32-core 3.2GHz Xeon.

The game was played with Chinese rules but without the usual compensation to white for the handicap stones.

2009-02-09

2009-02-13
Chun-Hsun Chou {Jun-Xun Zhou *) 9p MoGo 9×9 even 4-0 SGF SGF SGF SGF Taiwan Open 2009 MoGo was running on 640 cores of the Huygens supercomputer in Amsterdam.

In every case where each player won at last one game, the human lost the first game played and won the rest. This may be because of experience gained in the first game, or because of techniques learned from discussions with the other players.

Shih Chin 2p 1-1 SGF SGF
Li-Chen Chien 1p 2-0 SGF SGF
Cheng-Wen Dong 7d 1-1 SGF SGF
Biing-Shiun Luoh 6d 1-1 SGF SGF
Shang-Rong Tsai 6d 1-0 SGF
Shi-Jim Yen 6d 0-2 SGF SGF
Chun-Hsun Chou 9p 19×19 7 stones 3-1 SGF SGF SGF SGF
Shih Chin 2p 7 stones 2-0 SGF SGF
Li-Chen Chien 1p 6 stones 1-1 SGF SGF
Cheng-Wen Dong 7d 3 stones 1-0 SGF
       2008
2008-12-14Kaori Aoba 4pCrazy Stone19×19 7 stones0-1 SGF2nd UEC Cup Crazy Stone was running on an eight-core PC. The game has been described as "very beautiful".
2008-12-11Motoki Noguchi 7dMoGo9×9even 2-2 SGF SGF SGF SGF Event organised by Clermont-Ferrand Go Club Time limits were 30 minutes each, sudden death. MoGo was running on 640 cores of the Huygens cluster.
2008-09-27Zhou Junxun 9pMoGo19×197 stones1-0 SGF World 9×9 Computer Go Championship Zhou Junxun's account of these games says, in brief "In the first game, at move 11, I was shocked to realise I already had a lost position. But rather than panicking, I took time to devise a hamate. It worked. The second 9×9 game was relatively easy. In the 19×19 game I knew I would win by move 30."
9×9even2-0 SGF SGF
2008-09-21Myungwan Kim 8pMoGo19×197 stones2-0 SGF SGF Cotsen Open MoGo used an 200 node supercomputer with 4 cores per node. For the first game the time limit was 15 minutes each sudden death; MoGo chased a ladder in which its opponent had an extra liberty from moves 39 to 73, making it very easy for him to win.

For the second game, the time limits were 90 minutes each sudden death.

Three 3ks, a 4k, and a 5kMany Faces of Goeven2-3These games were played in the main event of the tournament, with all the other players being human. Many Faces of Go was running on a 2-core system.
2008-09-04Kaori Aoba 4pCrazy Stone19×198 stones 0-1 SGF FIT2008 conference on the Shonan campus of Keio University, Kanagawa prefecture, Japan. Crazy Stone was running on an 8-core system.
O Meien (Wan Mingwan) 9p9×9no komi0-1 SGF
2008-08-07Myungwan Kim 8pMoGo19×199 stones0-1 SGF 2008 US Go Congress. MoGo used an 800 node supercomputer with 4 cores per node.
2008-08-06Xiao Ai-Lin 1pLeela9×9no komi1-0 SGF 2008 European Go Congress.
2008-03-23Catalin Taranu 5pMoGo19×199 stones1-0 SGF Paris Go Tournament Played partly on a 256-core system and partly on a 4-core.
2008-03-229×9even2-1 SGF SGF SGF MoGo's 256-core platform only worked for one of these three games, the one which it won.
       2007
2007-06Guo Juan 5pMoGo9×9no komi?2-1 Played at 2007 Computer Olympiad Blitz.
       2006
2006-09 Kocsis and Szepesvári's seminal paper Bandit based Monte-Carlo Planning was published this September.
       2005
2005-10-30A 13-year-old "1k" (about 6d?)Many Faces of Go19×1915 stones0-1 SGF Played at the Gifu Challenge CupThe boy was nervous, playing on stage in front of an audience, with simultaneous pro commentary, so he played to avoid blunders, and gave up a lot of points in the endgame.

reported to me by David Fotland
       2003
2003-05Jean-loup Gailly 5kGo++19×199 stones0-1 SGF real date may be earlier
       2001
2001-03-03A Korean insei (1-gup, "1-kyu professional")Many Faces of Go19×1915 stones0-1 SGF Played at the Garosu Cupreported to me by David Fotland
       2000
2000-11-06three Taiwanese inseisWulu19×199 stones3-0 Ing Prize
       1999
1999-11three Taiwanese inseis, strong 6dsGo4++19×199 stones3-0 Ing Prize
       1998
1998-11-22three Taiwanese inseisMany Faces of Go19×199 stones 3-0 SGF SGF SGF Ing Prize
1998-07-18Jean-loup Gailly 5kHandTalk19×1917 stones1-0 SGF real date may be earlier
1998-08-05Martin Müller 5dThe Many Faces of Go19×1929 stones1-0 SGF
1998-02-23Jean-loup Gailly 5kGo++19×1920 stones1-0 SGF real date may be earlier
       1997
1997-11-12three Taiwanese inseisHandTalk19×1911 stones1-2 Ing Prize HandTalk won 250,000 Taiwanese dollars from the Ing Foundation
1997-08-??Janice Kim 1pMany Faces of Go9×93 stones0-1 SGF Played at the North American (or U.S.) Go CongressJanice Kim was playing a set of simultaneous games at the time, and not giving this her full concentration.

Reported to me by David Fotland.
1997-07HandTalk 19×1925 stones1-0 AAAI-97
       1996
1996-11-17three young playersHandTalk19×1911 stones3-0 Ing Prize
       1995
1995-11-19a 9-year-old and two 10-year-old youth championsHandTalk19×1913 stones1-2 Ing Prize HandTalk won 200,000 Taiwanese dollars from the Ing Foundation
15 stones1-2 HandTalk won 150,000 Taiwanese dollars from the Ing Foundation
       1994
1994-11-05three top youth playersGo Intellect19×1915 stones3-0 Ing Prize
       1993
1993-11three 12-year-old 3dsHandTalk19×1915 stones2-1 Ing Prize
       1991
1991-11three young 5dsGoliath19×1915 stones3-0 Ing Prize
17 stones0-3 Goliath won 100,000 Taiwanese dollars from the Ing Foundation
       1989
1989-11-1116-year-old 6dGoliath19×1917 stones1-0 Ing Prize
       1986
1986-11-11unknowna program by Du Gui Chong9×9even1-0 Ing event

* Chun-Hsun Chou, Chun-Shung Chou and Jun-Xun Zhou are the same person. He is from Taiwan, and "Chun-Hsun Chou" or "Chou Chunhson" is the Taiwanese rendering of his name into latin characters, while "Jun-Xun Zhou" is the mainland Chinese rendering.

The Ing Prize

From 1985 through 2000, Ing Foundation (properly The Ing Chang-Ki Wei-Chi Education Foundation) sponsored an annual computer Go tournament, preferring to hold it on the eleventh day of the eleventh month, for numerological reasons favoured by Mr. Ing. This was originally called the World Computer Go Championship, but came to be known as the Ing Cup.

The Ing Foundation arranged generous prizes for these events. At least for the later events, the first-placed program receieved 200,000 Taiwanese dollars, with 40,000 for second place, and 20,000 for third place. (Between 1992 and 2000 the Taiwanese dollar fell in value from 26 to the US dollar, to 31 to the US dollar, so the first prize was worth about US$ 7,000.)

As well as providing this prize money, the Ing Foundation also arranged for the winning program to play a challenege match, of a game against each of a set of inseis. These were trainee professional Go players from Taiwan, typically in their early teens, with strengths corresponding to about amateur 6-dan European. If the program won a majority of its games in this match, it won a further prize. There was one prize (to be won once only, then the next prize in the list would be competed for) at each of a series of handicaps, as listed:

HandicapVictory margin requiredPrize, Taiwanese dollarsstatus
"16 Moves", =17 stones2 of 3100,000Won by Goliath in 1991
"14 Moves", =15 stones2 of 3150,000Won by HandTalk in 1995
"12 Moves", =13 stones2 of 3200,000Won by HandTalk in 1995
"10 Moves", =11 stones2 of 3250,000Won by HandTalk in 1997
"8 Moves", =9 stones2 of 3400,000never won
"7 Moves", =8 stones2 of 3550,000 
"6 Moves", =7 stones2 of 3700,000 
"5 Moves", =6 stones2 of 3850,000 
"4 Moves", =5 stones2 of 31,000,000 
"3 Moves", =4 stones2 of 32,000,000 
"2 Moves", =3 stones2 of 35,000,000 
"1 Move", =2 stones2 of 310,000,000 
"First play", =no komi3 of 520,000,000 
Even4 of 740,000,000 

The final prize, for demonstrating a strength close to professional 1-dan, was worth over US $1,000,000, so this became known as the "million-dollar prize".

There has been some confusion about the handicaps involved. I believe that the above table is correct, and that some of the other versions which have been published are mistaken. In this table "17 stones" means, as is usual in the West, that Black places the first 17 stones, and then White places a stone. This differs from the usage of the Ing Foundation, which describes this as a 16-move handicap, as 16 stones are placed before the alternation is begun by Black. This terminology may be the cause of some of the confusion.


If you have ideas for other information that could usefully be here, please email the webmaster, Nick Wedd.

Past computer-computer results
future events
computer-go.info
HK's calendar
Computer Go article on Wikipedia