pychess-variants VS Fairy-Stockfish

I've just tried a chess variant called Shako, where there are two new pieces: Elephant and Cannon. Wowww, the Cannon piece is amazing ! This piece moves and captures like the Cannon in Chinese chess. It's not too powerful, but quite useful. It brings a new perspective and different to attack to the game. The Cannon creates many 'discovered attacks' situations. It's awesome. If you haven't played it yet, give it a try: https://www.pychess.org/

I want to share https://www.pychess.org/ as well - while it doesn't have hexagonal chess, it has tons of variants and you can play both against the computer (that runs fairy stockfish - an engine that works with many chess variants) and against human players

My current favorite is https://www.pychess.org/variants/empire - it's been designed to be wildly asymmetric but actually the balance between the two armies are more even than regular chess' white vs black pieces.

My problem is that there is no literature, no published theory on most chess variants. At least I can't find any theory on empire ehess, and thus I have to develop strategies on my own - but that's how I played chess 20 years ago and it was awful.

The best part of chess IMO is the social aspect of having an evolving theory (or as other games call it, an evolving metagame), and the way we can learn theory to be able to quickly assess who is winning at a given position (a highly nontrivial task), or assess whether a move is "natural" or not.

Is it on https://www.pychess.org yet?

It's already on https://www.pychess.org under the name "Placement Chess".

Fairy-Stockfish supports bughouse https://github.com/fairy-stockfish/Fairy-Stockfish/releases

Try Fairy Stockfish. Not as strong as regular stockfish but still should be more than enough.

I actually already implemented it in Fairy-Stockfish, although as a separate branch: https://github.com/ianfab/Fairy-Stockfish/tree/duck. As we currently still reuse the chess evaluation until we have a first NN(UE) it isn't on top human level yet, but move ordering already is decent, so once we have a proper evaluation it should do much better.

implement your favorite Anarchy Chess rules with fairy-stockfish (https://github.com/ianfab/Fairy-Stockfish), which lichess already uses for the variants it hosts, so it should be easy™ to make your cloned lichess accept it

Oh, apparently an experimental fork of fairy-stockfish with duck chess support was published yesterday! https://github.com/ianfab/Fairy-Stockfish/tree/duck

To my knowledge it isn't. Some of the small-sized shogi variants, like Dobutsu, Micro, or Kyoto shogi, can either be (practically) weakly solved using engine search or even strongly using tablebases in the case of dobutsu shogi. However, for minishogi neither of those have yielded a definitive result yet to my knowledge. Nevertheless, deep engine searches and games from high level engines strongly indicate that gote, i.e., the second player, wins with perfect play.

You might want to take a look at - https://github.com/ianfab/Fairy-Stockfish

But, if you have a variant which the free Fairy Stockfish or ChessV (Alt link) can not play, Zillions can probably implement it fairly easily. Just this year, I created a Chess variant using an unusual tiling of squares and triangles. The only engine which can play this variant is Zillions.