Normalizing the eval

[Michael Sherwin]

Code: Select all

void Normalize() { 
  s32 sq; 
  s32 highWN; 
  s32 highBN; 
  s32 highWB; 
  s32 highBB; 
  s32 highWR; 
  s32 highBR; 
  s32 highWQ; 
  s32 highBQ; 
  s32 limit;

  highWN = -INFINITY; 
  highBN = -INFINITY; 
  highWB = -INFINITY; 
  highBB = -INFINITY; 
  highWR = -INFINITY; 
  highBR = -INFINITY; 
  highWQ = -INFINITY; 
  highBQ = -INFINITY;

  limit = 90;

  for&#40;sq = A1; sq <= H8; sq++) &#123; 
    if&#40;wKnightTbl&#91;sq&#93; > highWN&#41; highWN = wKnightTbl&#91;sq&#93;; 
    if&#40;bKnightTbl&#91;sq&#93; > highBN&#41; highBN = bKnightTbl&#91;sq&#93;; 
    if&#40;wBishopTbl&#91;sq&#93; > highWB&#41; highWB = wBishopTbl&#91;sq&#93;; 
    if&#40;bBishopTbl&#91;sq&#93; > highBB&#41; highBB = bBishopTbl&#91;sq&#93;; 
    if&#40;wRookTbl&#91;sq&#93;   > highWR&#41; highWR = wRookTbl&#91;sq&#93;; 
    if&#40;bRookTbl&#91;sq&#93;   > highBR&#41; highBR = bRookTbl&#91;sq&#93;; 
    if&#40;wQueenTbl&#91;sq&#93;  > highWQ&#41; highWQ = wQueenTbl&#91;sq&#93;; 
    if&#40;bQueenTbl&#91;sq&#93;  > highBQ&#41; highBQ = bQueenTbl&#91;sq&#93;; 
  &#125; 

  for&#40;sq = A1; sq <= H8; sq++) &#123; 
    if&#40;highWN > limit&#41; wKnightTbl&#91;sq&#93; *= &#40;limit / highWN&#41;; 
    if&#40;highBN > limit&#41; bKnightTbl&#91;sq&#93; *= &#40;limit / highBN&#41;; 
    if&#40;highWB > limit&#41; wBishopTbl&#91;sq&#93; *= &#40;limit / highWB&#41;; 
    if&#40;highBB > limit&#41; bBishopTbl&#91;sq&#93; *= &#40;limit / highBB&#41;; 
    if&#40;highWR > limit&#41; wRookTbl&#91;sq&#93;   *= &#40;limit / highWR&#41;; 
    if&#40;highBR > limit&#41; bRookTbl&#91;sq&#93;   *= &#40;limit / highBR&#41;; 
    if&#40;highWQ > limit&#41; wQueenTbl&#91;sq&#93;  *= &#40;limit / highWQ&#41;; 
    if&#40;highBQ > limit&#41; bQueenTbl&#91;sq&#93;  *= &#40;limit / highBQ&#41;; 
  &#125;
&#125;

I have limit set to 90 centi pawns for all pieces. My question is, do I need different limits for different piece types, game stages, circumstances and what is the best guess for those values. Thanks!

I should add that the eval for each piece is all accumulated into the piece square tables.

Code: Select all

    // Knight Tables
    moves = knight[sq] & ~aPieces;
    trgts = knight[sq] & (bRooks | bQueens | bKings);
    wKnightTbl[sq] = PopCnt(moves) * 4 + PopCnt(trgts) * 8;
    wKnightTbl[sq] += ((cenRow + cenRow) * (cenCol + cenCol) + row * 3);
    if(cenCol == 0) wKnightTbl[sq] -= 20;
    g = evalHistTblg[WN][sq]; n = evalHistTbln[WN][sq];
    wKnightTbl[sq] += (n > 99 ? g/n : 0);

[Michael Sherwin]

First quick test of above idea. RomiChess vs Monolith. Best previous result Romi 53%. Early results for this test.

RomiChess64P3o - Monolith_03_x64 : 21.0/31 16-5-10 (=010=1101=1110011==1==1111=11==) 68% +131


Will see if it holds up.