A Shogi Champion Turns to Chess

FFriedel writes "Michael Jordan tried it with baseball, and it, like, didn't work out too well for him. But what about a professional Shogi champion switching to chess? Yoshiharu Habu, one of the most gifted players in the history of the ancient Japanese game, has taken a casual interest in chess - and already reached IM strength. He is currently playing in a tournament in Paris, where chess grandmaster Joel Lautier interviewed him." Shogi is a very odd game if you're used to chess. Most of the pieces have biases toward forward motion, and when you capture an enemy piece, you can bring it back into play for your side.

Comments

Shogi and Go

[nagora]

Both games have a strange aspect which westerners and programmers find tricky to handle: there are times when a game is over which the rules do not define but leave up to the players to agree. Basically the rules say "once the game can't be won, its over" without defining "can't". Programming this is difficult.

FWIW, shogi is far more fun than chess and more interesting too.

TWW

Re:It's like ya know.....

[Anonymous Coward]

The submitter's section is quoted straight from the introduction of the article. (you would know that if you'd read the article, the submitter == the author)

The author is Frederic Friedel, who also happens to be the guy behind much of chessbase.com, I think. He is probably not a native English speaker; he speaks several languages (judging from Google and chessbase research), including German, English, and Spanish. He is also an expert on computer chess. (and a very good chess player, having conversed with Garry Kasparov on things like the Brains of the World puzzle)

This is just an aside. I believe that linguistically speaking, the 'like,' bit was a sly dig at Michael Jordan and US culture, i.e. he used 'valley girl' lingo to express the 'oh well it didn't work out, so what' impression that he may have received from Jordan's failed baseball career. (and contrasted that with the undisputed Shogi grandmaster who has achieved International Master status barely a decade after teaching himself chess by reading a book about it)

Now, much of the linguistic implication outlined above is relatively speculative. But it suggests that he knew what he was talking above, and that it had some 'deeper' meaning.

Besides, making fun of someone who is a fairly high-level chess player and who speaks several languages well enough to converse and report in them, simply because he used some 'funny' phrasing, is kind of silly. Don't you think?

(posting anonymously to preserve my precious karma)

ASBTA

Computers still struggling ...

[arvindn]

Lautier: Are computers a threat for Shogi ?

Habu: In mating problems, called Tsume Shogi, the computer is already superior to the best players. In normal games, however, the computer is still far from the professional level. Its level can be compared to a 4-dan among amateurs [approximately 2300 strength in chess Elo terms. The first dan among professionals starts after the amateur 6-dan. To get a rough idea, the best Shogi players in the world, including Mr Habu, have a ranking of professional 9-dan - JL].

This is one example of the prevailing sad state of affairs of the performance of AI in games. The best chessplaying programs are those which use brute force search and little else. The fact that they can beat world champions tells us little except that the effective branching factor in chess is small. In games like go and shogi where the branching factor is much higher, long-term strategy counts much more, and brute-force is relatively useless, computers are nowhere near the best humans.

Another example: As early as 1962 Samuels wrote a checker playing program which could learn from its previous games and beat reasonably strong humans. After that there has been virtually no progress in game strategy; all the improvement has been in hardware speed. Indeed, it wasn't until 1994 that the first wold-champion-beating checker player, "Chinook" was written. This is an amazingly slow rate of progress compared to other areas of computer science/technology.

Its a shame, considering that game playing is thought to be one of the easiest problem domains for AI.

Re:Computers still struggling ...

[Anonymous Coward]

Top-flight backgammon has been revolutionized by the use of neural networks, starting in about 1990. Computer backgammon had suffered from the large branching factor problem, since each ply of lookahead involves an average of 15-20 ways to play a roll, times 21 possible dice rolls, for a typical branching factor of 300-400. The problem still exists, of course, but a neural network makes an outstanding static evaluator. The modern backgammon programs play a good game with no lookahead at all, and play as well as any human when the neural net is combined with a 2 ply lookahead.

Shogi can improve your chess

[atabotix]

A few years ago an US IM, Larry Kaufman, learned shogi, became good at it, and reported that it improved his chess. I recall the article in Chess Life & Review but it's hard to find a good web ref to this. This is a stab at it: http://www.msoworld.com/mindzine/news/orient/shogi / auf.html

Use Yahoo IM and Google? Try YIMGoogle