I was wondering, is it possible to program a chess computer like, indestructible? I mean, if you set a computer to like, 2750 strength, is it even possible to win from that computer? I mean, it analyses every possible move and won't make any mistakes while humen tend to make inaccuracies which the computer will make use of to it's fullest. I think the max you can get out of that is a draw by not making any single mistakes yourself.

But then again, what genius is able to give a computer a rating strength of 2750?

Also, what determines the strength of a computer?

I suppose it's the analysing time, would be something like;

• 0.2 seconds = 1200
• 0.4 seconds = 1500
• 0.6 seconds = 1800

Like that?

Then strength of a computer depends on:

Chess engine (software) that you are using.

RAM available

Length allowed for analysis.

I'm half way through reading Behind Deep Blue by Feng-Hsiung Hsu. He led the team who built the computer that beat Kasparov in a match in 1997. He says it wasn't really 'man versus machine' - it was man as performer (Kasparov) versus man as toolmaker (the Deep Blue team). It's a very well-written book, and has to be, because most readers will be like me, and won't understand either the chess or the computing.

I noticed on playchess.com that the highest rated computer "players" were those that had the most processing power i.e. a 2.4 GHz quad core running Rybka 2.3mp could outperform a 2.4 GHz dual core running Rybka 2.3mp. So along with RAM and software, architecture plays an important role, too.

crisy, if you have time, I recommend that you watch this documentary:

http://video.google.com/videoplay?docid=4895271762581045075&ei=T5OLSZSQJYyG-wGin9nSBg&q=kasparov+deep+blue&hl=en&emb=1

This is about the match, however from the perspective of Kasparov. It is very intriguing, and I thoroughly enjoyed it.

Now that's interesting. I could have never even thought of the faster the computer, the better the chess software.

So, if you have a slow computer like, an 800 mhz processor, it means chess software is useless? After all it's by far not fast enough to do any better analysis of games than you do!

It doesn't mean that the analysis is useless, it just means that it will take a longer time to reach a certain depth than a faster computer would. 

Ah, a faster computer would not have better analysis than a slow one, it would just take less time. Pretty logical, feel dumb for asking.

The processing time would probably be around O(n^k), or exponential time, because the processing of moves would likely take a tree format (i.e. look through all moves, if a move looks good, look through all followup moves, etc.).

So it might take 0.2 seconds to play at 1200 level, but maybe 2 seconds for 1800 level, 30 seconds for 2400 level, and so on.