The auto-tuned PST tables show quite acceptable results, but I see a problem with them in the middlegame due to the position of the king (in the center, short and long castle). For example, in PESTO and Chili, the tendency towards short rocking clearly prevails, which is bad when the rocking is long. Multiple castling must also be considered.
The idea is to create 16 (4 * 4) tables that take into account all the standard positions of kings in the middlegame.
The query is mainly about the lithander, it has working mechanisms for creating such tables and for quickly testing the hypothesis, what is expected to be improved from their application.
// the positions of the kings are taken from the table:
int wking[64]={
0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,
3,3,3,0,0,0,2,2,
3,3,3,0,1,0,2,2,
};
int bking[64]={
3,3,3,0,1,0,2,2,
3,3,3,0,0,0,2,2,
0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,
};
// milti mg PST:
mgpst[4][4][5 or 6?][64]; // mgpst[wking[wk_sq]][bking[bk_sq]][piece][square]
What do you think?