Go back
are there tricks to beating computer chess program

are there tricks to beating computer chess program

Only Chess

Vote Up
Vote Down

Worth checking out the book 'A Psychiatrist Matches Wits with Fritz' - written by an 1800-2000 strength player who consistently defeated Fritz 5 and 6 using what he calls barrage positions (from the Colle or Birds, where there is a long-term attack that is beyond the programs horizon).

There's a comprehensive review on Chessbase: http://www.chessbase.com/newsdetail.asp?newsid=316

Vote Up
Vote Down

Originally posted by Osse
Worth checking out the book 'A Psychiatrist Matches Wits with Fritz' - written by an 1800-2000 strength player who consistently defeated Fritz 5 and 6 using what he calls barrage positions (from the Colle or Birds, where there is a long-term attack that is beyond the programs horizon).

There's a comprehensive review on Chessbase: http://www.chessbase.com/newsdetail.asp?newsid=316
Until I joined this I played only on an electronic chess game (rated approx 2200) as I was not able to find any decent opposition.After 5 years of playing it I can beat it on it's highest setting at about 80% of the time.But this means going into a room and locking the wife and kids out.I find that opening and middle play should be directed towards the end game,so slow build up of pawns etc.I think that programmes are written based on greed and their tactic seems to be about getting into a position where you are a piece down,pawn or knight to a bishop then finishing you off.You need to look as far forward as you con an not be drawn into exchanges early on as you will find that if you start trading pieces all of a sudden you will find your self in a very difficult position.Beware that I have found that these tactics can prove awkward in otb chess as although you can develop a strong game you will find that medium to storng players can catch you off guard as it is a different kind of game, use it but be on your guard

Vote Up
Vote Down

Makes you wonder - if it`s that easy why are they banned on here? After all, what`s the difference between using a database or a computer?
Especially as they seem so easy to defeat.

Vote Up
Vote Down

Originally posted by davfra
Makes you wonder - if it`s that easy why are they banned on here? After all, what`s the difference between using a database or a computer?
Especially as they seem so easy to defeat.
I'm not saying it is that easy,but after playing only electronic chess for 5yrs you start to get e feel for what to expect from this position and that.I have played very few games with my electronic chess as black for instance and mybe after another 5yrs I might be able to expect a good sucess rate.I do not have any computer games like Fritz etc and would probably loose much more often. All I'm saying is they have their short comings and this is what I've found.Of course I can always turn it off and start again if I'm loosing and I can take back, if I make a mistake,But when you're playing properly you can't do that.It's good for training but what posiible good is it if you win games using one,you may as well just have the rhp computer chess site instead.

Vote Up
Vote Down

I think the following is a good example of a computer program's lack of understanding of positional concepts:

consider this position:

  • 8
  • a
  • 7
  • b
  • 6
  • c
  • 5
  • d
  • 4
  • e
  • 3
  • f
  • 2
  • g
  • 1
  • h



If you play white in this position against, for example, Pocket Fritz 2, all you have to do is shuffle your King back and forth between d2 and e2... chances are that Black will merely shuffle it's Rooks and Bishop around harmlessly, and you will probably be able to draw by the 50 move rule, if you don't hit a 3-fold rep first...

If you play with Black in this situation, you would probably eventually figure out that the only way to win would be to break through the pawn chain...and the only way to do THAT would be to sacrifice at least one Rook, and possibly two...and then promote a pawn... (the Bishop can not be sacrificed in this way, because it's a light-square bishop, and the pawns are all on dark squares).

Playing this position on Pocket Fritz 2, the program "scores" this position as about a 14.9 point advantage for Black... yet, while playing black, Fritz does not want to give up a rook, and therefore, it will merely shuffle it's Rooks and it's Bishop behind it's own pawns, until 50 moves have been played. It's truly an amazing sight to see...

I'm sure other programs will do the same, although PF2 is the only program I've played this postion on, since PF2 is the strongest of the 4 chess programs I have on my PDA...

Try it, and see if you have any programs that understand that a rook sacrifice is required to win in this position... I haven't tried this on Deep Fritz 8, Shredder or Chessmaster 10th, all of which I have...

And, for what it's worth, when I've beaten the higher rated personalities on Chessmaster (for me, that's 1500+, and i've only won a few of those), I usually did it by keeping the position as closed as possible... sometimes, against a computer, the program simply will not 'know' what to do in a closed position...

And I would agree that exchanging pieces with a strong software program is asking for trouble...at least, that's been my experience...



Vote Up
Vote Down

Originally posted by TheBloop
I think the following is a good example of a computer program's lack of understanding of positional concepts:

If you play with Black in this situation, you would probably eventually figure out that the only way to win would be to break through the pawn chain...and the only way to do THAT would be to sacrifice at least one Rook, and possibly two...and th ...[text shortened]... ed in this way, because it's a light-square bishop, and the pawns are all on dark squares).
Perhaps I'm stupid, or my eyesight is failing me, but I can't see any possible way that black can force a way through that pawn chain. That is without White being a willing participant in the process.

It seems to me that if white wants a draw here, it is pretty much guaranteed.

--tmetzler

Vote Up
Vote Down

Couple of other thing about the above position I posted:

If you're playing white in this position, and a computer program ever does offer a rook sacrifice, you would, as white, decline the sacrifice and continue to shuffle your king around until you get a draw...


If you're playing black, and you offer the rook sacrifice, the computer , as white, will take the rook every time...

I just played this position with Deep Fritz 8... as Black, it merely shuffled it's rooks and bishop around... as white, it immediately took the rook when I moved the rooks to the a - file and moved one rook to a5... the pawn immediately took it, which allowed me to take the pawn with my other rook, then advance my b-pawn, forcing another exchange, and then black can break through with the King and Bishop and force a pawn promotion.

Vote Up
Vote Down

Originally posted by TheBloop
Couple of other thing about the above position I posted:

If you're playing white in this position, and a computer program ever does offer a rook sacrifice, you would, as white, decline the sacrifice and continue to shuffle your king around until you get a draw...


If you're playing black, and you offer the rook sacrifice, the computer , as white, w ...[text shortened]... hange, and then black can break through with the King and Bishop and force a pawn promotion.

I played my electronic chess game last night and won.I did open e4,e5
then I spent most of my time deveoping the pawns(sorry I don't make notes)but making sure they were backed up by the bigger pieces.Then the game castled 0.0 at about 20 odd moves into the game so I decided to build an attack slowly (as a previous head on assault from a similar position failed and the game was drawn).In essance the conclusion was decided by overall understanding of the position,I sacrificed three pieces to gain positional advantage and win the game and during this time the game
played three moves which showed it was not able to undrestand the overall position,moves that a good player would have made and the game would have been drawn.It seems to not understand the advantage to be gained by sacrificing pieces for positional advantage and certainly never does it itself.I'm sorry for my vagueness but as I said I don't make notes and I am not familiar with all the opening terms even if I play them.As I said I have played this same game for 5 yrs probably three times a week so I guess that gives me a bit of a head start and I have found that since joining rhp I it is much harder to get to grips with the variables thrown up when playing different poeple.

2 edits
Vote Up
Vote Down

Originally posted by tmetzler
Perhaps I'm stupid, or my eyesight is failing me, but I can't see any possible way that black can force a way through that pawn chain. That is without White being a willing participant in the process.

It seems to me that if white wants a draw here, it is pretty much guaranteed.

--tmetzler
No, you're right...a draw is guaranteed for white, just by shuffling the King behind the pawns...

But if you play this postion against DF 8 or PFritz2, and you move your rook to a5, the computer program will grab the rook, which will then allow your b-pawn to advance, force another pawn exchange and then, supported by your King and your other rook (which may need to be sacrificed, depending on the computer's response) you can break thru and eventually queen a pawn...

It IS impossible for black to force it's way through the pawn chain, without White being a willing participant...you are absolutely correct about that... my point (since I was talking about chess software's relative lack of understanding of positional concepts) was that DF8 or PF2 WILL BE a willing particpant, because they'll grab the rook when you post it on a5...

Unless white takes the bait (and the rook), the game will end in a draw, either by 50 move or 3 fold rep.


All my example really was was one way of showing how, in playing against a computer form this position, you can either (1) force a draw while playing white, in spite of Black overwhelming material advantage, or (2) force the win while playing black, by tricking white into grabbing a major piece (i.e. your rook), which it will do every time. There's no way that DF8 or P Fritz 2, as white, will leave that rook alone, even though that's the only way to avoid losing the game.

Vote Up
Vote Down

are there tricks to beating computer chess program


Switch it off.

Vote Up
Vote Down

Originally posted by XanthosNZ
(and the position after 25. e4 appeals a great deal).

Understatement. That is...flabbergatsing.

Vote Up
Vote Down

`If you're playing white in this position, and a computer program ever does offer a rook sacrifice, you would, as white, decline the sacrifice and continue to shuffle your king around until you get a draw... `

Exactly - So you are at fault, not the computer. The computer knows that there is no way it can force you to accept the rook sacrifice.
The computer doesn`t work on the probability that you will make a mistake. That`s why in some ways the computer is cleverer than us.
I think that you have demonstrated that the computer has a better grasp of the position than yourself!
No offence intended.

Vote Up
Vote Down

Originally posted by davfra
`If you're playing white in this position, and a computer program ever does offer a rook sacrifice, you would, as white, decline the sacrifice and continue to shuffle your king around until you get a draw... `

Exactly - So you are at fault, not the computer. The computer knows that there is no way it can force you to accept the rook sacrifice.
The comput ...[text shortened]... ed that the computer has a better grasp of the position than yourself!
No offence intended.

No offense taken...

The position I gave in this example was taken from an "anti-computer" web site... I don't have that link anymore, I had it on a computer which I no longer use... the author of that site was merely illustrating a point about how most software will react to certain types of closed positions...

The author said that the reason that the computer will not offer a rook sacrifice (even though that is the ONLY way as black to ultimately win the game), provided white takes the rook) is because it will not give up the material advantage it has with two rooks and a bishop vs. a lone king...

The other point is that, if you play black, the computer WILL take the rook if you move it to a5...it doesn't 'understand' that taking the rook is the only way it will lose the game...it just wants to grab the material.
That's all the author was saying...there is a lack of understanding of certain postional concepts in most commercial software...

So if I am playing Black, and the computer is playing white, it does NOT have a better grasp of the situation than I do, because it will take the rook, and that will allow me to win the game...the computer won't play for the draw, even though it's far behind in material in this situation.
It's evaluation function will tell it to grab the rook, which it then will do. It doesn't understand that all it needs to do is shuffle it's king around, and there would be no way for black to get at him.

1 edit
Vote Up
Vote Down

Vote Up
Vote Down

i may be misinformed about how a computer "thinks" about chess, but i thought that the blunt approach was to consider several possible moves and their potential outcomes- so why isn't a computer able to see that the rook sacrifice would lead to a break through and ultimately a win? or how does the computer miss that Not taking the rook is the only way to ensure a draw? i realize that it' working on the idea that "rook is better than pawn- take rook now" kinda mentality, but can't it calculate to only a few movs ahead when it becomes obvious that a pawn will be promoted?

Cookies help us deliver our Services. By using our Services or clicking I agree, you agree to our use of cookies. Learn More.