Comments/Ratings for a Single Item
OK thanks. The problem was that the improvised way in which the UI passed the moves to the AI only worked correctly for Pawn moves (fD or fA). It always passed only a single e.p. square, and if a forward step was made, this was rounded to the square of origin. On sideway orthogonal moves it even crashed the code through division by 0.
This should be fixed now. (Refresh browser cache!)
I am still not entirely happy with it, as when you pass through a check by a non-Pawn it doesn't actually perform the winning capture, but just continues with another move. So while playing you have to be careful not to make illegal moves with the royal Queens. Fortunately the UI will always gray out those. I will try to fix that later.
Steps to reporduce: put queen on b6, put pawn on a5, start the AI and move the queen to b5
- Qdb6 aa5 2. Qb5 axb6
OK, that is not really e.p. capture. I could not reproduce that, but there definitely is something wrong, as I sometimes get JavaScript errors when I move the Queen out of check. I will look into it.
Very interesting. Having been working on Parton's works recently, I find this idea quite good. Parton had explored having two K in Twin Chess / Chess Tweedle with a different principle. Both K of one side were both royal, so checkmating the first one was a victory. Here that idea is completely renewed. :=)
I think you misunderstood. AI checked my queen with pawn, I moved out of harms way (to the side of the pawn) and the pawn captured my queen on a square it wasnt on.
I have however noticed an issue with the interactive diagram. When I tried palying with the queens, the AI took my queen en passant, I assume that is not intentional.
The rules say the Queens cannot pass through check. So if you stuck to the rules nothing could have captured you en passant. The capture is the diagram's way to enforce that rule, just as it would capture the royal in the normal way when you illegally step into check, or move away a pinned piece. It captures the King, and declares "I win!".
I did not test how well balanced it is; I could not find a way to easily modify Fairy-Max to support the 'not-passin-through check' rule.
Interesting variant, i love asymetric variants with elegant rules. The question is how balanced it is, did you balance it with engine?
I have however noticed an issue with the interactive diagram. When I tried palying with the queens, the AI took my queen en passant, I assume that is not intentional.
In Spartan Chess having a spare King is worth about 4 Pawns (halfway between minor and Bishop). Unlike in Pink Chess it is not allowed to leave both Kings under attack there ('duplecheck'), but I do not expect that to affect the value significantly.
This is a really interesting idea! I have to hand it to you, you keep coming up with interesting new ideas. And, given how many variants there are now, truly innovative ideas that are actually fun to play are few and far between.
I'd like to run some tests on which side is stronger, but that will be difficult. For one thing, we'd need to assign a value to the king. How much is having the second kind worth? We would have to run tests to try to determine that before we could have any confidence in tests to decide which side is stronger.
11 comments displayed
Permalink to the exact comments currently displayed.
can you castle through attacked square if you still have another king