2

Opening traps can sometimes lead to quick wins, but fail against opponents who are prepared or have time to calculate. Thus...

Question: Is there software which generates novel chess opening traps?

I didn't easily find anything via Googling chess opening trap generator. Perhaps there's some inherent limitation which makes it not worthwhile or even impossible to do so.

Improve this question
3
  • 7
    I think the main problem is that "trap" is a very human concept. There must be a move that seems "obvious" except for a refutation that's "harder to see". But those are not engine concepts, every move is equally easy to see for an engine. So I wouldn't know how to build such a tool.
    – RemcoGerlich
    Feb 1, 2021 at 8:56
  • 2
    @RemcoGerlich Maia Chess might make this possible.
    – Federico Poloni
    Feb 1, 2021 at 9:07
  • It seems TrappyBeowulf was built with that idea, but I've just found the code is quite old and the compilation doesn't properly work: github.com/vollmerm/TrappyBeowulf
    – emdio
    Feb 3, 2021 at 13:55

1 Answer 1

Reset to default
1

I am sorry, but you can not create any novel chess opening traps because all were already discovered in a certain range of moves. Also, I want to mention @RemcoGerlich's comment that even if there was an undiscovered opening trap then a computer couldn't recognize that it is a trap.

Instead, you could try to use an engine to look up moves that lose but are difficult to comprehend by humans.

Improve this answer
3
  • 1
    I'm wondering: The "Maia" chess engine is trained to behave "like a human" at a certain grading range, and playing against a low-rated version, I've definitely had it walk into opening "traps" against me (Trying to hold on to the captured pawn in a QGA-type opening where doing so is dubious). So maybe with such an engine you could mine for "trappy" openings?
    – Lagerbaer
    Feb 3, 2021 at 21:23
  • @Lagerbaer Nice comment. Obviously chess engines can walk into opening traps if they are on a low level, but they can not differentiate whether this is also difficult for human are if the engine just blundered. However, I understand what you mean. For example, an engine could look at situations in which there are only very little possibilities to maintain a good position.
    – Hacker
    Feb 3, 2021 at 21:35
  • Hi Hacker, what you say is certainly true for an engine like Stockfish or Leela. But this Maia engine was specifically trained to not play "the best move" but to play the move that "a human would play", and there's different versions of it for different rating ranges. That gives much more useful information about whether a certain move is "hard to find" for a human.
    – Lagerbaer
    Feb 5, 2021 at 15:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.