The chess computer

Are there any specific weaknesses chess computers have that us humans can utilize? And no I'm not asking for ways to disconnect the power or causing a short circuit lol

I think misevaluating positions by putting too much emphasis on material and not enough on positional considerations is one. I've seen cases of very imbalanced gambits (like the Halloween) being used to great effect against engines by exploiting this tendency.

Thanks. People will laugh at this but I am trying to defeat the easy computer xD I'll look into some obscure gambits to use as a weapon

Chess computers Are Tactical and Can see Everything! But just because they see everything doesn't mean they can stop it. Positional Chess is the only way to win. However, if your Playing Houdini, Rybka or some other Top Level Engine. Then its Impossible because their programming is specifically designed to be pure Math. Along with Endgame TableBases and Opening Books. Your only real Chance is in the Middle Game. But Top Level Engines are Incredible and can see the problem in the position Immediately. They Basically think 2+2=4. So this is the Best move! Its as simple as that. Only Super GMs could ever stand a Chance. Out of 10 games, they would probably win 3 and Draw a few. But if your talking about Chess Computers in General(The one your computer comes with or something) Then just play Positional Chess! Its your only Chance. Good Luck!

chess engines do not always recognise closed positions as draws.....

just close the position against an engine and  u may shake hands with engine...

computers are perfect at tactics but bad at planning. but it's not something that can be exploited easily.

Splendid. I'll try these.

Endgames outside of tablebases result in ludicrous engine analyses. I've put Fritz 13 on deep analysis and it makes incredible endgame conclusions that are totally wrong. In particular engines will assess a position as an equal = endgame because the "lost position" is 30 moves down the track beyond even their horizon.

At the level of anyone in this thread, inc me? I doubt it. If you can make me a liar I'll be more than happy for you, it's not like I'm rooting for your PC or I'm a traitor to my species! 

What sort of TL(s) do you plan on playing vs the PC? What chess program do  you have? What sort of PC is it running on? Not that any of this really matters (except for your TLs)  but it wouldn't  hurt to tell people anyway. 

The easiest way to equalize your chances vs a PC is to configure it to play at about your rating if there is such an option. If not limit it's "thinking" ie depth of ply search etc to levels that are close to your own. You can also have it give you "time odds" but that alone probably isn't enough vs today's software. You can also set up positions (such as openings or endgames you want to learn) to practice and train with where your side has the advantage and see if that  helps much or not. I wouldn't be too worried about my results vs the PC as long as I felt I was learning something or getting better in general from playing "it."

All this is probably a moot point unless you're already familar with how chess software works and how to configure it. If not it's time to RTFM and you'll figure out how to create weaknesses in the software (maybe) I'm thinking in terms of Fritz and other programs with lots of features and options to regulate their play with.

NimzoRoy wrote:

kamuimaru wrote:

Are there any specific weaknesses chess computers have that us humans can utilize? 

At the level of anyone in this thread, inc me? I doubt it. If you can make me a liar I'll be more than happy for you, it's not like I'm rooting for your PC or I'm a traitor to my species! 

What sort of TL(s) do you plan on playing vs the PC? What chess program do  you have? What sort of PC is it running on? Not that any of this really matters (except for your TLs)  but it wouldn't  hurt to tell people anyway. 

The easiest way to equalize your chances vs a PC is to configure it to play at about your rating if there is such an option. If not limit it's "thinking" ie depth of ply search etc to levels that are close to your own. You can also have it give you "time odds" but that alone probably isn't enough vs today's software. You can also set up positions (such as openings or endgames you want to learn) to practice and train with where your side has the advantage and see if that  helps much or not. I wouldn't be too worried about my results vs the PC as long as I felt I was learning something or getting better in general from playing "it."

All this is probably a moot point unless you're already familar with how chess software works and how to configure it. If not it's time to RTFM and you'll figure out how to create weaknesses in the software (maybe) I'm thinking in terms of Fritz and other programs with lots of features and options to regulate their play with.

What's a TL? And I am talking about the chess.com live comp ;;

Engines don't always understand endgames well enough because they have to calculate many moves forward. By the way, I wrote another article here on General Discussions about using engines.

FM Järvenpää

I'm not going to use an engine.

People have different views about how nowadays engines manage to win top human players. First of all we have to mention on what engine we are talking about. I have Houdini 3 so I mention few things about that. Although in past Kasparov managed to win Deep Blue the situation is quite different now. Carlsen vs Houdini 10 game match would probably end up 2 draws and 8 losses for Carlsen. It's already too difficult to beat the best engines that are far away above 3000 level. I believe there are some results of the matches between new engines and top human players that tell the same story that they almost never lose to humans.

If someone wants to beat an engine ( no matter is it Houdini 3 or some version of Fritz or Rybka) then let's not forget the good advice to go for closed positions! I am sure they have complications because in closed positions it's not as much about tactics but more of strategical planning which for engines is only based on tactics.

I managed to draw houdini at full strength like 2 or 3 times by creating closed positions from the opening and locking the center without any tactics for the computer to break through. One time I managed to trade to be a bishop up in the endgame but unable to convert it to a win.

Does that mean that I can never beat the good computers because I basically think the same way a computer does, only more selectively and not at millions of nodes a second?