DiscoCheck 5.2
It scores +58 elo against DiscoCheck 5.1, at fast time control 6"+0.05". Probably +40 elo in rating list conditions (ie. against unrelated opponents and at much longer time control).
Important changes
Update move sorting heuristics only when alpha is raised. This stupid bug was quite a big elo drain.
Passed pawn tuning: push away more aggressively the enemy king from our passed pawns.
Simplify Pawn Piece-Square Table
Introduce material imbalance formula, based on counting minor pieces. Re-tune material values in the presence of the new material imbalance.
Do not distinguish node type for reduction (PV, Cut, All are always reduced the same way).
null search only when depth >= 2. And re-CLOP null search reduction with depth >= 2 precondition.
Prune threat refutations, like any other moves.
Improve LMR transition (between search and qsearch).
Introduce triple reductions.
Do not double reduce bad checks.
Never prune the first move.
Update move sorting heuristics on TT pruning (History, Refutation, Killers).
http://www.open-chess.org/download/file.php?id=588
It scores +58 elo against DiscoCheck 5.1, at fast time control 6"+0.05". Probably +40 elo in rating list conditions (ie. against unrelated opponents and at much longer time control).
Important changes
Update move sorting heuristics only when alpha is raised. This stupid bug was quite a big elo drain.
Passed pawn tuning: push away more aggressively the enemy king from our passed pawns.
Simplify Pawn Piece-Square Table
Introduce material imbalance formula, based on counting minor pieces. Re-tune material values in the presence of the new material imbalance.
Do not distinguish node type for reduction (PV, Cut, All are always reduced the same way).
null search only when depth >= 2. And re-CLOP null search reduction with depth >= 2 precondition.
Prune threat refutations, like any other moves.
Improve LMR transition (between search and qsearch).
Introduce triple reductions.
Do not double reduce bad checks.
Never prune the first move.
Update move sorting heuristics on TT pruning (History, Refutation, Killers).
http://www.open-chess.org/download/file.php?id=588
