Recent computer progress in Shogi.

[rubin427]

I really don't know much about shogi, or the state of computer play of shogi, but I found this brief write-up of both a recent challenge sent to the Japan Shogi Association, and a follow-up regarding the results very interesting.

The Challenge

The Outcome

The posts were made on a language blog - so if someone is very interested in the topic, they may want to verify the news against another source (which I have not taken the time to do, mosty becuase I don't speak japanese).

[hirohiigo]

I can't access the articles because I don't have an account. But you can see the kifu of the game here:

http://wiki.optus.nu/shogi/index.php?cm ... &kid=72657

I'm pretty impressed that computer programs have caught up with Joryu-Oushou Ichiyo Shimizu, who for a long time has been one of the best lady professionals. I would be interested to see it play an actual pro again in the future.

[hyperpape]

I have no familiarity with Shogi--I assume that the difference in computer play is because it's more complicated than Chess, not just because of lack of attention?

[Joaz Banbeck]

I was trying to get a quick grasp of how shogi is played, and I came across this on wikipedia:

"On 14 October 2005, the Japan Shogi Association banned professional shogi players from competing against a computer. The Japan Shogi Association said the rule is to preserve the dignity of its professionals, and to make the most of computer shogi as a potential business opportunity. The ban prevents the rating of computers relative to professional players."

( It had a link to the English version of Japan Times which said little more. )

[pwaldron]

An English writeup at http://www.newscientist.com/blogs/short ... g-mov.html

[Suji]

This is impressive.

I assume that the difference in computer play is because it's more complicated than Chess, not just because of lack of attention?

Correct. The dropping of pieces makes Shogi more complicated than Chess. Endgames with few pieces on the board are rare. The dropping of the pieces makes sure that the effective branching factor stays relatively high.

[hirohiigo]

Here's an English article about the game on Mainichi Daily News: http://mdn.mainichi.jp/mdnnews/news/201 ... 2000c.html

The last time a computer played against a professional was in 2007 when the Bonanza engine played against Ryu-ou Akira Watanabe. Watanabe defeated Bonanza then.

I assume that the difference in computer play is because it's more complicated than Chess, not just because of lack of attention?

Creating shogi engines is a very big field of research in Japan. As was pointed out earlier, the ability to drop pieces to virtually any square on the board multiplies the game complexity greatly compared to chess.