Explore >> Select a destination
|
You are here 
|
knhash.in | ||
| | | | |
edjohnsonwilliams.co.uk
|
|
| | | | | Looking at some modified versions of chess | |
| | | | |
proceedings.neurips.cc
|
|
| | | | | ||
| | | | |
zserge.com
|
|
| | | | | How to write a simple chess playing program in one day with a few lines of code | |
| | | | |
mbuffett.com
|
|
| | | I wrote a post yesterday about how to compress chess moves. I was pretty happy with my hand-rolled compression scheme, which achieved a per-move size of 9.5 bits. I almost just rolled with it, but I've been thoroughly nerd-sniped here, so I took another crack at it. To know the board state or not One of the main considerations in compressing chess moves is how much chess-specific computation you want to do. One of the most optimal forms of compression is actually to evaluate the position, use an engine like Stockfish, and then store the index of the move. This isn't really a viable method though, the CPU cost is insane. But that's just to highlight what one end of the spectrum looks like. | ||
