Nowadays there are so many chess engines and they are so good that there are very little chances of improvements. Google Deepmind AlphaZero, Stockfish, LeelaChessZero and Komodo are few of the top engines who took chess to a different level. These levels are beyond human capabilities. With neural networks and large chessbook databases these engines defined how chess should be played. But somehow in all these upgrades the fun is loosing. We can't understand their moves. That somehow reduces our enthusiam in understanding these engines. But whenever I see a chess varient I feel more entertained. So I'm trying to build different chess varients.
- Indian Chess(where pawn can move only one step even on its first move)
- No-castling(the king can't castle anymore)
- Stalemate=Win
- Pawn-back(pawns can move one step backward also)
- 2-steps-any-time(pawns can move 1 or 2 steps anytime)
- Pawn-sideways
- Own-piece-capture(one can capture his own pieces)