Mig 
Greengard's ChessNinja.com

Topalov on Comps

| Permalink | 15 comments

A fluffy Reuters item on CNN.com has Veselin Topalov talking about chess and computers. Interesting timing since he's not doing anything with them in public right now. Or is he? Perhaps he was sought out by a reporter under the influence of the recent New Yorker item on computer chess.

"A human, a world champion or a top grandmaster at his best, should still be better, but you achieve this condition only a few days a year," he said. "In a long game, people still have a chance, even if it's not much of one."

Not exactly the tough talk you might expect were Topalov angling for a big-money match with Hydra. I agree with the comment in the article that says GMs often don't take computer games seriously. For most they are occasional exhibitions and treated as such. Kramnik and Kasparov prepared hard for matches against Fritz and Junior in 2002-03 and could only draw. Kramnik and Kasparov! I doubt lesser players think they can survive against stronger machines with little or no preparation. They simply don't much care. If they lose, they're almost supposed to and there is no loss of Elo, reputation, or pride.

15 Comments

Computers are calculating machines, and so that is all they should use. No more opening books - at least not when playing humans. Besides, the first time Hydra's son beats the World Champ with 1. a4, it'll be more exciting than when he grinds him (or her!) down in some overbooked line of the Sicilian.

Topalov was the toughest opponent against computers in Amsterdam so it is not going to be one-way traffic as in Adams-Hydra. But I don't fancy ANYONE against Hydra...

BTW did you know that Morozevitch defaulted today because he ...overslept?!!

In the case of Hydra the monster usually gets out of the book as early as move 6 so it's not going make a lot of difference.

"A fluffy Reuters item"? I can assure everyone I had nothing to do with this piece.

Sakaev was obviously talking through his hat when he said Moro overslept. Now everyone repeats that as if it was a fact, which it isn't!
We are still in the dark why there was a no-show.

Volkov played 26. g3?? instantly losing! 26. Kg1 was winning. Nice to see one move blunders from on high!
Lots of laughs, laugh out loud, etc.

Whether the opening books are taken away or not, Hydra's evaluation function would find 1.e4 because that is the strongest move in terms of mobility and various other factors. Indeed, anyone is free to download Crafty, use it without opening books, and witness the phenomenon. None of today's sophisticated programs would ever play 1.a4.

"Sakaev was obviously talking through his hat when he said Moro overslept" ...according to 64 Moro defaulted in bed!

According to chesspro.ru, Morozevich was 1 hour 12 minutes late - those 12 minutes made him forfeit, since if you're late by an hour or less, you still get to play.

Also, don't the games start at 15:00 local time? Must have been one heck of a night if he overslept for a game starting at 3PM.

About opening books: Whenever I play Chessmaster 9000, full strenght but no opening book, the beast does not beat me easily (I am not a GM, not even close to it). So I conclude that computers donīt play good chess. Is it a matter of time? I wonder if you allow Hydra, Fritz, Schredder, Chessmaster and others play against each other, a large number of times (say, for the next fifty years), accumulate experience, that is, build an opening and end game database, maybe then they will play better than top GMīs.

Mig anything about the Max Dlugy story we should know thats not already posted on chessbase.com?

I heard Morozevich took a page from Kotov and is working on a new book. It's called "Sleep like a Grandmaster"

The perfect match with a computer is one where equals conditions is met. That's include these at least:

1. No opening book (or at least the same for both). Remember Kasparov lost the 6th game in the opening with Deep Blue?

2. Control with the tablebases. The computer must not use it at all.

If we want to test any computer program we need that it resolve all its positions with its own calculations. Otherwise the human player plays against his memory or knowledge of opening or ending theory, more than the machine.

"no loss of Elo, reputation, or pride" when a top player looses against a computer : Are you serious Mig? Not even talking about money (Adams has made 10000 $ where he could expect reasonably 60000 $), not even talking about money Adams is just playing like a zombie since his inhuman match against Hydra...

Mickey's the last person to be hurt by such a loss. He's a rock. He simply didn't play well in San Luis. Statistically speaking, someone had to finish with a negative score...

I expect to talk to Max after his return. If he has anything to add, I'll be happy to add it. The little I know in addition to the news stories is personal stuff it will be up to him to say.

Twitter Updates

    Follow me on Twitter

     

    Archives

    About this Entry

    This page contains a single entry by Mig published on December 21, 2005 1:54 AM.

    2005 Russian Superfinal was the previous entry in this blog.

    Freed GM is the next entry in this blog.

    Find recent content on the main index or look in the archives to find all content.