Calling the castling rule in Chess 960 "tricky" is putting it mildly. As a game developer, I’d honestly prefer to drop it entirely. But officially, it boils down to this: after castling, the king and rook end up on the exact same final squares as they would in traditional chess. All the standard conditions still apply — neither piece can have moved previously, and you can’t castle out of, through, or into check.
Picture this: a rook on a1 and a king on a2. After castling: they land on d1 and c1. That means both pieces jump in the exact same direction rather than leaping over against each other. As a player, it goes insanely against chess instincts: you’re used to that beautiful, automatic, physical feel of hiding the King behind a Rook, but Chess 960 just breaks physics when the king jumps out of the castle rather than jumping in. It
feels like gravity turned off. And as a developer, coding this logic is pure agony.
I only wonder why the Bobby didn't just propose swapping the king and the rook for castling — it's way more logical than what we have now. Alright na, I think I'll try it (don't shoot)
Full disclosure: Fischer Random castling is not here on this site just yet, partly on principle (omg!) — it will be the moment I'll gather enough courage to hack into the core engine to refactor it for the 960's stupid castle rule. But hey, I've already cooked that Fischer Random for joy as the primary goal, it delivers ton of fun, it definitely
kicks in!