Don't treat king safety differently in AnalysisMode
authorJoona Kiiski <joona.kiiski@gmail.com>
Sun, 7 Apr 2013 08:23:52 +0000 (09:23 +0100)
committerMarco Costalba <mcostalba@gmail.com>
Tue, 9 Apr 2013 21:29:58 +0000 (23:29 +0200)
Rationale:

- Current settings seem to make engine *significantly* weaker in analysis mode.
- In practice this setting only has effect when king safety scores are high.
- Even in analysis mode its far more important to know if one side is getting mated,
rather than get evaluation correct with 1cp accuracy.

No functional change

src/evaluate.cpp

index 3d9a3112fdc47167297d512d346d1c9bb344c622..6990a71d2bbb306d945cfedebc94c659e1497475 100644 (file)
@@ -289,14 +289,6 @@ namespace Eval {
     Weights[KingDangerUs]   = weight_option("Cowardice", "Cowardice", WeightsInternal[KingDangerUs]);
     Weights[KingDangerThem] = weight_option("Aggressiveness", "Aggressiveness", WeightsInternal[KingDangerThem]);
 
-    // King safety is asymmetrical. Our king danger level is weighted by
-    // "Cowardice" UCI parameter, instead the opponent one by "Aggressiveness".
-    // If running in analysis mode, make sure we use symmetrical king safety. We
-    // do this by replacing both Weights[kingDangerUs] and Weights[kingDangerThem]
-    // by their average.
-    if (Options["UCI_AnalyseMode"])
-        Weights[KingDangerUs] = Weights[KingDangerThem] = (Weights[KingDangerUs] + Weights[KingDangerThem]) / 2;
-
     const int MaxSlope = 30;
     const int Peak = 1280;