Deep Blue vs Deep Blue

[Wes Janson]

In infinite/infinite number of games, the starting computer would win, or else there would be a draw (I don't remember which it is). If the starting computer wins, then it would always win. If it's always a draw, then it's always a draw. I recall that Kasparov did win some of the games, implying that it is possible to win against Deep Blue, but I don't know the circumstances. In all likelihood, it would result in a stalemate every time.

[Off Center]

Most computers use Dynamic Programing to play games like chess. Dynamic Programing is a method of solving larger problems by combining the solution to smaller problem groups. Basically, what happens is that the computer assigns a point value to the next move based on what occurs that turn. For instance, a lost piece will be assigned a -1, no exchange a 0, and taking a piece a +1 (obviously, this is a simplified approach; points will vary based on which pieces are exchanged and their location on the board). The computer does this for each piece for several turns based on the level of difficulty at which the program is set. The highest score resulting score will dictate the next move. This is why, if you are playing chess on the computer, it will take the computer longer to move at higher difficulty levels. This, of course, introduces an element of randomness if two resulting scores are equal for a given move, as the computer must make a decision as to which piece to move.

That being said, if I remember correctly, white (goes first) has the advantage in chess because. If white plays a perfect game, they will win.

[MrWallen]

Not sure on this, but I believe that the Deep Blue was noy just programmed to play chess, but to play chess against the current reigning grandmaster (at the time), Kasparov I think?

Anyway, I'm pretty sure it was programmed with solely him in mind.

[Dryden]

Deep Blue was designed to play reactionary to chess game types. It was designed to beat an opponent by predicting against a style of play, not the opponent's specific board positions. Kasparov beat Deep Blue both in the '96 match and in two of the '97 games by using one opening, then switching to a totally different mid-game and end-game. Of the 6 games in the '97 match, Deep Blue won 3, Kasparov won 2, and one ended in a draw.

The '97 edition of Deep Blue was an order of magnitude more powerful than it's '96 predecessor, and still managed to lose two games.

The real problem is whether Deep Blue would always compute board positions in the exact same order, which it should not due to it's predictive modeling logic.

The mathematical problem is this:

Deep Blue can examine 200,000,000 board positions per second.

By comparison, Garry Kasparov can examine 3 board positions per second.

There are a total of 10^120 board positions in all of chess. To give you an idea of the magnitude of this number, that's more than the number of nanoseconds since the Big Bang, and more than the number of atoms that make up the known universe.

No computer can figure out every possible board combination, period. No computer can even figure all playable combinations from a single specific board position at given points in the game if the number of pieces exceeds x, where x would be variable given the unique movement characteristics of the available pieces. On a 3 minute classical chess clock, chess playing computers have to tree-predict the most likely playable scenarios and combinations, 5 to 10 moves ahead, using only the current board position. Deep Blue can (or could?) do about 200 billion positions in three minutes.

By that, I would think that every game would be random, since there are more possible moves at mid-game than even the most powerful supercomputer can analyze. The game should become random even if every opening were the same, since at the very least two computers would recognize each other's openings as being different in their branch-prediction.

As an example, even in the end-game, when there are so few pieces left that one would think Deep Blue knows every possible outcome ... well, let's just remember Kasparov still managed to beat it, several times.

[AssassN]

Originally posted by Dryden
well, let's just remember Kasparov still managed to beat it, several times.

That's insane. I can't even beat my home computer.

[sharpshooter1286]

man is smarter than machine, he built it..machines are meant for a purpose, a job, and thats not to be dominant over man

[Ultimator]

Originally posted by sharpshooter1286
man is smarter than machine, he built it..machines are meant for a purpose, a job, and thats not to be dominant over man

What kind of backwards-arse philosophy is that? Let's see you try to analyze 200,000,000 chess board positions per second.

Yeah that's what I thought.

Here's something just to keep on topic: I suck at chess.

Thanks, Bye.

[MrWallen]

Originally posted by Ultimator
What kind of backwards-arse philosophy is that? Let's see you try to analyze 200,000,000 chess board positions per second.

Yeah that's what I thought.

Here's something just to keep on topic: I suck at chess.

Thanks, Bye.

Just because a computer can do it FASTER doesn't mean it's more intelligent.

[sharpshooter1286]

exactly mr wallen, u understood what i meant

ultimator, if a rat can navigate a maze faster than a blind man, i suppose hes more intelligent to? because thats the kind of thing ur implying that the faster u do something the smarter u are

[Ultimator]

You're still diverting from your original statement. Or maybe I misunderstood you.

You're saying that man built machine and assigned it a job, and that job is not to be dominant over man?

What?

[MrWallen]

Its job was to be better than man, not more intelligent.
An axe is better at chopping wood that a human hand, does that make it more intelligent? No.
Deep Blue is better at playing chess than a human, does that make it more intelligent? Nope.

[sharpshooter1286]

ultimator now i think ur disagreeing just to act smart, and to be sarcastic...grow up. while ur completely missing what i mean, u said machine was smarter than man correct? am i correct in saying that would mean machines are dominant over man. man did not create them so that something would be smarter than them, machines are a tool used to do something not overpower man and how does my statement not make sense?

You're saying that man built machine and assigned it a job, and that job is not to be dominant over man?

What?
-i said man built machine for a purpose, and the purpose wasnt to overpower man. how is that not clear? do u have the reading level of a 4th grader or something?