Chess Software Clasification

Maybe a software Like Fritz can be clasified as agresive, pasive, positional, or something like that.

It came to my mind since fritz is going to play Kramnik, So can a machine be clasified into a category?

Any coments, or knowledge about it?

I guess so. I have Junior10 and it is very aggressive and loves to attack.

All computers are essentially tactics monsters in nature. So they are all "attackers". Albeit some are designed to be more "strategic", this is not the same thing as a human style.

If you think about what they’re doing this makes sense. Basically, they are just looking at millions of possible variations and applying a simplistic evaluation algorithm and finding the position that maximizes the function.

So a computer will tend to optimize factors which are easily detectable (i.e. space and material); and are best when the most important factors are short term in nature. This is why they are great tactically. Tactics tend to be short term ideas involving elements which are easily evaluated by the computer.

For Kramnik to beat Fritz10, he will have to create situations that are long term in nature, or involve elements the evaluation function can’t identify. This is how he defeated Fritz 7, he forced an endgame situation that he knew how to win, but the computer could not understand (and it wasn't short term enough to find by brute force).

He also will have to stay away from ultra sharp positions, in which he has no hope of matching the computer's near perfect tactics. When tactics do arise, he will have to at least match the computer.

Originally posted by Spacetime
All computers are essentially tactics monsters in nature. So they are all "attackers". Albeit some are designed to be more "strategic", this is not the same thing as a human style.

If you think about what they’re doing this makes sense. Basically, they are just looking at millions of possible variations and applying a simplistic evaluation algorith ...[text shortened]... near perfect tactics. When tactics do arise, he will have to at least match the computer.

Regarding the Kramnik vrs Fritz match. I wonder what the computers propensity for accepting a draw is? If they reach 40 moves into a game that is clearly drawn but could take some time to finish, it would be advantageous for the machine to play it out as fatigue is a factor for Kramnik, but not for the machine.

And yes, certain programs are tailored to certain styles. I believe that either HIARCS or Rybka come the closest to impersonating human play.

Originally posted by sonhouse
When is that match scheduled?