How resource-intensive would it be for the server analysis to use the tablebases in the explorer? I have seen positions where the server analysis and local analysis run to high depths evaluate a position as a draw (0), but according to the tablebase is a forced win for one side. As I play the tablebase moves into the game, the local engine evaluation favours the winning side more and more until a checkmate score is announced (e.g mate in 12), but the "actual checkmate length" following the tablebase moves is longer.
TLDR, it would be really cool if the server analysis could consult a tablebase in order to improve evaluation.
TLDR, it would be really cool if the server analysis could consult a tablebase in order to improve evaluation.