computer chess

i'm probably just over-thinking here, but maybe somebody can help me out.. somewhere down the line, i picked up the impression that if a computer plays itself at chess (with the proper strategy's in its database) that the white team will win every time, because white always moves first and therefore has that slight mathematical advantage..

does anybody know for certain, is that correct?

hi notepad,

By the way I also beleive what you are telling is correct... Are you meaning a situation where both white & black players controled by computer itself?
Then your argument becomes really true. Computer will not think more than stratergies stored on the database. So, then everytime the game goes as same as the previous times... Hopefully, White player has the slight advantage, and will win the contest!!!

yes, i'm talkin about computer vs computer both using the same database of strategy's.

I think you are preparing for the next coding contest.. Matt might have appointed you as the admin on charge of it..
Why don't u just say Script vs Script

Though they use the same strategy what we are talking about is the game of chess where moves are not the same.. In that case at somepoint the strategy for the last move made by the opponent may lead black players to take charge..
I think so..

To expand on what nazly said, I don't think the white advantage would be true. Yes, they may have the same database of strategies, but that does not mean that white and black will take the same moves. A database of strategies for a computer chess player would not involve clear cut move paths. It would be based upon the current situation of the board, so white and black would use different strategies.

It all comes down to the strategies that in in the database....

Not just the strategies to, but computer logic ... If you play chess against a computer playing its best, it will recognise youre moves and make the best match against it, so each time will be picking best against each other, so each game could go either way...

well, according to the research i've done, a computer gaming program isn't really considered "perfect" unless it can always win or at least reach a stalemate so long as it makes the first move. for example:

so ASSUMING you had the perfect pc, then this "perfect" chess program could be possible no? my friend argues something along the lines of "adaptive neural net algorithm" but i don't know what the heck he's talking about (postalcow would probably get it).

MY REASON, for asking, is not for a contest.. but rather that i have a pretty good reputation at finding flaws within every movie that i watch (i'm quite the movie goer) and last night i bought "wargames" on dvd. for those of you don't know, it's an old-school hacker movie from the 80's that makes quite a few references to defcon.

WARNING:
do not read any further if you have not seen the movie "wargames" cause i am about to spoil it
END WARNING

okay so the whole plot really of this super self-learning machine is to teach itself that sometimes it's okay to give up, because you can't always win.. WELL, if this chess theory is at all legit (not to mention there were plenty of other games on the computer that could have had databases small enough to fit this theory, with the exception of tic-tac-toe of course). you see what i'm getting at? ..the computer should have already known that you can't always win, why did it need tic-tac-toe to learn that? THIS TEARS THE WHOLE PLOT APART. and then after it learns it can't win, it says "that was a strange game.. how about a nice game of chess?" but uhh.. mmhrmm..

Ive seen the film, and uve spoiled it lol, but i guess theirs no perfect pc, so no chess game, or other game can win ALL the time ... but think about it, playing itself at chess, wud be like u playing urself at chess trying your hardest ...

there is (at least) one major drawback in your thinking.

imagine if you will you playing with your self (and you are a perfect player, and you have in your brain all that previous games on record, etc ;).

now, when you come to the move, let say half the game down the line, you have an option of two good moves.

you calculated every possible position after those moves, and you see that they are both as good. so, you chose one at random.

and you lose the game. now, the next time you play, when you are in the same position, you will remember the last game, and you will chuse a different move, right?


also, if you play the first move, it might be better if you opponent plays something, but worse if he plays something else. i know you are going to say that the opponent (you) is just as perfect, so he will always play the same (perfect) move in that situation, but, just because of that, he could try to trick you and play that other move that is almoust as good as that one..


or... i don't know.. i don't think it would ever be possible to build that perfect computer. i mean, for that kind of computer, playing all combinations of tic-tac-toe would be over in under a millisecond, so we see that computer (was it joshua) from the film wasn't in fact perfe ct, so, he could still see that chess could be interesting to him ;)

I see where youre coming from ... But think about it, if you're playing with yourself and using a strategy thats always worked for you, that strategy isnt gonna work if you use it with both sides, so you improvise, but a computer cant do that, so it will end up changing strategy and the you cant immediatly say 'whites will win' coz u dont know that ...

but it is possible to never lose at tic-tac-toe, as it is with connect-4.. so why should chess be any different? it'd just take a lot longer to calculate is all, it's mathmatics.

i got the impression that joshua was a more powerful machine than i think even exists today, i mean it was practically A.I., which is why i think that it should have been able to handle this chess theory.

Its possible to never lose at any game, but with tic tac toe its possible to trick the opponent into the wrong move, same with connect 4, and definately easier to do in chess ...

But to trick would infer a human emotion - and although Joshua displayed some human charactistics (mostly due to it's AI I guess) - to be 'tricked' does not constituet the 'perfect' game of chess in the way that connect-four has been solved.

In defence of Joshua - if it was AL (Artificial Life - this is where the adaptive neural net bit comes in) then it would have to learn that chess can always be won - as it had not built up all its data by that point in time.

You adapt to your situation - so you begin playing chess thinking your never going to win, then you start getting a bit better, then you start learning about openings ect, and with the help of a super computer chip and, oh say, about a couple of centuries, you work out how to always win. So I think the film is okay!

J

If it's the computer playing against itself, then tricking is not really going to work. That is if the computer is 'perfect'. Because you see, if it were 'perfect' it would know the trap, and therefore not fall for it.