Double Impact of Gain tables
authorJoona Kiiski <joona.kiiski@gmail.com>
Sun, 31 Mar 2013 17:35:45 +0000 (18:35 +0100)
committerMarco Costalba <mcostalba@gmail.com>
Wed, 3 Apr 2013 17:14:59 +0000 (19:14 +0200)
Very unorthodox idea.

After 16000 games at 60"+0.05
ELO: 3.14 +-3.4 (95%) LOS: 96.6%
Total: 13407 W: 2278 L: 2157 D: 8972

bench: 4705335

src/search.cpp

index ed029b4c4a08e6d15b3ef28f2b4b246b04ad74d9..5ff23ad0fd894791c7b5e8ac625246c40343e1e5 100644 (file)
@@ -875,7 +875,7 @@ split_point_start: // At split points actual search starts from here
           // but fixing this made program slightly weaker.
           Depth predictedDepth = newDepth - reduction<PvNode>(depth, moveCount);
           futilityValue =  ss->staticEval + ss->evalMargin + futility_margin(predictedDepth, moveCount)
-                         + Gain[pos.piece_moved(move)][to_sq(move)];
+                         + 2 * Gain[pos.piece_moved(move)][to_sq(move)];
 
           if (futilityValue < beta)
           {