News for hackers. Stuff that matters.
Re:The neural network of the Stockfish chess engine by EvgeniyZh on Thursday January 14, 2021 @02:27AM (#25773273)
Current divp uses very unbalanced openings for engines to be able to score. Using starting position/more optimal openings inevitably leads to draw. So it's actually exactly the opposite -- the advantage is so low and draw tendencies are so strong that we need to force position where white has advantage to be able to compare two strong engines

There are no comments on this comment yet.