[Computer-go] What about the exhibition games?

Hideki Kato hideki_katoh at ybb.ne.jp
Sun Feb 16 17:05:59 PST 2014


All game records of the 1st Denosen are attached (byoyomi was added 
in the time fields).

Hiroshi's notation of Japanese names may confuse readers.  Emura and 
Ozawa are the familiy names and Kikou is usually spelled Kiko, like 
Tokyo (never written as Toukyou :-).

All games used Japanese rules and 6.5 komi.  The time setting 
for 13x13 was 30 min and 30 seconds byoyomi.  For 19x19, 60 and 60 
but Ozawa used less than half.

Emura 7d (9d on Tygem; Japanese 7d is a special rank; almost a title for 
top-level amateurs) is sooooo strong.  Zen had no chances at all.  Maybe 
one or two (no-komi or 2-stones) handicap was necessary.  His fuseki 
was better than Zen's (well, Zen's fuseki is often said 1-dan-level :) 
and the depth and correctness of the forecasting for the fight at the 
bottom left corner in the second game clearely overtook Zen's.  He 
forecast the result (W would won the fight) at about move 23 (!) while 
Zen noticed at move 34 and the winrate dropped to 30%.  Please note 
that this position is very complicated and the order of the moves is 
very important.  The game was over.  Zen continued playing, trying 
critical moves, but Emura replied in very cool.  

For the first game, Emura played so-called slow moves in the middle of 
the game and Zen looked have a chance (indeed, the winrate got 60% once) 
when Zen could play three big yose moves, but, at least I think, Zen had 
no chance at all as well.

The last game was very interesting as Zen preferred territory (!).  
Zen has played all public games in center-oriented style.  Even against 
Takemiya 9d, very famous with his cosmic style, Zen got cener but, 
this game, Zen got three corners.

Ozawa played with Zen-style, say, fourth-line fuseki, katatsuki for 
fourth-line stones, and o-moyo.  He is the first strong amateur 
played aginst Zen with Zen-style :).  So the games was relatively 
easier for Zen, I think.  No complex positions to simulate and 
evaluate.  Counting large territory is hard (with bigger error) for 
human but Zen.  The winrate was monotonicly increased throughout the 
game.   Ozawa notice he was behind in the middle of the game when 
Zen's winrate was about 60%.  The last ko fight may look dengerous but 
Zen counted it already.  Zen's lookahead showed only A18 was the 
effecive ko threat and Zen would ignore that and would win the game.  
Ozawa didin't realize that losing the ko killed his stones when he 
started the ko fight.  Zen had clear advantages in the two L&D, at the 
top left corner and right side.

If he played with other styles, such as territory oriented, the game 
might close.

Hideki

Hiroshi Yamashita: <2426080FDF8D409E90D6BB4401780AD0 at i3540>:
>Hi,
>
>> I am very curious. How did the exhibition games of
>> Zen (on 13x13 and 19x19) run today?
>> Are there already sgf available?
>
>Zen lost 2 games in 13x13. And won a game in 19x19.
>
>    Emura Kikou,   7d Ama - Zen   2-0
>    Ozawa Ichiro,         - Zen   0-1
>
>I don't have sgf files. 
>
>Photo news, Ozawa lost to computer.
>http://news.mynavi.jp/news/2014/02/16/108/
>
>Regards,
>Hiroshi Yamashita
>
>_______________________________________________
>Computer-go mailing list
>Computer-go at dvandva.org
>http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
-- 
Hideki Kato <mailto:hideki_katoh at ybb.ne.jp>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Game records.zip
Type: application/octet-stream
Size: 4690 bytes
Desc: not available
URL: <http://computer-go.org/pipermail/computer-go/attachments/20140217/e4ca31e4/attachment.obj>


More information about the Computer-go mailing list