3 ms·

Ah, I hadn't seen Adler's comment before. I did most of my work on this before he posted that comment. His scheme is probably more efficient than mine (he is

by Tommah 3y ago

Ah, I hadn't seen Adler's comment before. I did most of my work on this before he posted that comment. His scheme is probably more efficient than mine (he is Mark Adler, after all), but I'd have to run it on the same set of positions to know for sure. You figured out how I'm representing the position correctly, but I'm only storing one half-move per position, specifically the current player's best move. I use that for the "Show hint" feature.

lifthrasiir 3y ago
> I'm only storing one half-move per position, specifically the current player's best move. I use that for the "Show hint" feature. I think you are describing B = 1 case. When is the B != 1 case used then? I also wonder why the reference board seems to have an obvious mistake (rooks in place of bishops). Again, I hardly know any chess, so that might be my oversight and that does improve the compression for this particular data set.