My website, Chess Variants Training, just got a big update: we now have crazyhouse support!
You can solve crazyhouse puzzles here: chessvariants.training/Puzzle/Crazyhouse
Or you can play Crazyhouse960, Crazyhouse with a randomized back rank: chessvariants.training/Variant960
A special thanks to @causa-sui for building a crazyhouse puzzle generator that has created thousands of checkmate puzzles!
Register an account first, so that you can track your progress and won't be presented with the same puzzles each time you come to the site.
Take your time. In each position there is at least one forced mating line; however, only the fastest possible mate will be accepted. Slower mates will be rejected. (If there are two lines that mate equally fast, either one should be accepted.)
Have fun. :)
awesome!! i would love to play 960zh my name is mopperwezen on your site.
Incredible -- I was actually working on the exact same thing. Causa-Sui is the puzzle generator open-source?
Hi Flourish. The script I used can be viewed here gist.github.com/rpdelaney/a25f2bc42ec674b3901ce15a9355425b
Note that it's not completely bug free yet, but I was able to generate upwards of 5000 puzzles with it so far
One recommendation: list how many moves the mating sequence is for the zhouse tactics trainer (perhaps sort by mate sequence). Obvious second recommendation is to include less efficient mates, which will take considerable more programming effort. If I have some time, will check to see if I can write something up.
Showing the count of moves in the sequence is something we considered, but eventually I decided not to reveal it - people might not want to see it and find it themself. We could have a "hint" button, but if some people do that and others don't, the puzzle ratings might get somewhat disordered.
I have previously collected some material here, with similar intention. I checked them all with the engine. You can use them there
> Obvious second recommendation is to include less efficient mates, which will take considerable more programming effort.
I think this wouldn't require a huge amount of programming effort to adapt my script to do this. However, it would drastically increase the computation time, as the branching factor becomes unmanageable very quickly.
Also, I'm not clear that it is actually desirable in the first place. Sometimes second-best mates are just delayed versions of the best mates. After all, there is no time limit and the goal is to find the BEST move.