More aggressive LMR reductions.
Tested at different time controls:
- Tested with 1CPU 1+0, after 3000 games, result was +12 ELO
- Tested this with 4CPU 1+0 and got sth around 5-10 ELO increase
- Last one at long time control,after ~1000 games with 10+0 result is:
Orig - Mod: 491 - 520 (+10 elo)
A testing marathon by Joona for this important change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
for (int i = 1; i < 64; i++) // i == depth (OnePly = 1)
for (int j = 1; j < 64; j++) // j == moveNumber
{
for (int i = 1; i < 64; i++) // i == depth (OnePly = 1)
for (int j = 1; j < 64; j++) // j == moveNumber
{
- double pvRed = 0.5 + log(double(i)) * log(double(j)) / 6.0;
- double nonPVRed = 0.5 + log(double(i)) * log(double(j)) / 3.0;
+ double pvRed = log(double(i)) * log(double(j)) / 3.0;
+ double nonPVRed = log(double(i)) * log(double(j)) / 1.5;
PVReductionMatrix[i][j] = (int8_t) ( pvRed >= 1.0 ? floor( pvRed * int(OnePly)) : 0);
NonPVReductionMatrix[i][j] = (int8_t) (nonPVRed >= 1.0 ? floor(nonPVRed * int(OnePly)) : 0);
}
PVReductionMatrix[i][j] = (int8_t) ( pvRed >= 1.0 ? floor( pvRed * int(OnePly)) : 0);
NonPVReductionMatrix[i][j] = (int8_t) (nonPVRed >= 1.0 ? floor(nonPVRed * int(OnePly)) : 0);
}