I finally extended my book building framework to include iCE performance data for certain lines. In the past the database just consisted of lines played by strong programs. Lines that were played successfully more often received a higher chance to be selected in game play.
I now added win/draw/loss statistics that I collected when two similar versions of iCE played each other. For the starting position it now looks like this
FEN: rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1
EndGame Id: UNKNOWN
Played moves :
DataBase Content:
No Move Prob Played Perf % Wins Draws Losses
------------------------------------------------------------------------
1 e2e4 44 38.031 51.59 25.021 41.387 22.208
2 d2d4 41 36.392 53.22 24.494 40.629 19.076
3 c2c4 8 7.283 53.06 8.104 13.541 6.388
4 g1f3 6 5.019 52.99 8.173 13.948 6.463
5 g2g3 1 487 51.69 989 1.543 874
6 b2b3 0 298 51.93 250 417 216
7 b2b4 0 281 45.18 18 39 26
8 f2f4 0 275 45.80 116 259 161
9 b1c3 0 201 50.78 105 182 99
10 c2c3 0 132 75.00 7 4 1
11 e2e3 0 124 56.00 25 34 16
12 g2g4 0 13 0.00 0 0 0
13 d2d3 0 12 56.82 7 11 4
14 a2a3 0 3 42.65 6 17 11
15 h2h3 0 2 0.00 0 0 0
16 f2f3 0 1 0.00 0 0 0
17 a2a4 0 1 0.00 0 0 0
Book Content:
No Move Prob
--------------------
1 e2e4 44
2 d2d4 41
3 c2c4 8
4 g1f3 6
5 g2g3 1
When we dive into one line, eg the sicilian after 1. e4 c5 we come to
FEN: rnbqkbnr/pp1ppppp/8/2p5/4P3/8/PPPP1PPP/RNBQKBNR w KQkq - 0 2
EndGame Id: UNKNOWN
Played moves : e2e4 c7c5
DataBase Content:
No Move Prob Played Perf % Wins Draws Losses
------------------------------------------------------------------------
1 g1f3 86 12.764 51.21 9.042 14.326 8.279
2 b1c3 6 981 48.10 925 1.713 1.066
3 c2c3 6 935 52.21 678 1.295 566
4 g1e2 1 184 49.87 97 181 98
5 g2g3 1 88 45.18 27 49 38
6 f2f4 0 74 46.33 108 239 144
7 b2b3 0 71 46.18 63 116 83
8 d2d4 0 53 43.94 29 58 45
9 d2d3 0 51 46.33 92 157 119
10 c2c4 0 33 38.10 2 12 7
11 a2a3 0 17 44.64 8 9 11
12 b2b4 0 15 37.50 1 4 3
13 f1c4 0 10 25.00 0 1 1
14 b1a3 0 9 55.92 23 39 14
Book Content:
No Move Prob
--------------------
1 g1f3 93
2 c2c3 7
The database shows that the move 2. Nc3 is a common variation but iCE scores less than 50% with it, which is a bad score for white. So this move is now no longer copied into the book. If however iCE is playing black and the opponent selects this move it still knows how to continue
FEN: rnbqkbnr/pp1ppppp/8/2p5/4P3/2N5/PPPP1PPP/R1BQKBNR b KQkq - 1 2
EndGame Id: UNKNOWN
Played moves : e2e4 c7c5 b1c3
DataBase Content:
No Move Prob Played Perf % Wins Draws Losses
------------------------------------------------------------------------
1 b8c6 65 566 51.30 555 862 505
2 d7d6 22 192 52.18 383 601 326
3 e7e6 9 77 52.51 145 253 119
4 a7a6 2 18 48.60 40 93 45
5 g7g6 2 17 50.00 18 37 18
6 b7b6 0 2 0.00 0 0 0
7 g8f6 0 1 0.00 0 0 0
Book Content:
No Move Prob
--------------------
1 b8c6 65
2 d7d6 22
3 e7e6 9
4 a7a6 2
5 g7g6 2
I think it makes sense to build the books this way, it probably does make a difference ELO wise but anyway as a human I try to avoid lines I don't feel comfortably with also.
A little space to track the efforts spent on my hobby: Chess Programming
No comments:
Post a Comment