I added rook and queen on 7th rank evaluation to MadChess 2.0 Beta. The code awards a bonus for a rook or queen on the 7th rank if the enemy king is on the 8th rank (its 1st rank).
Also, I removed late move pruning. (When I implemented late move pruning in build 44 it added 39 Elo to the playing strength of MadChess 2.0 Beta.) I replaced it with more aggressive null and late move reductions, and more aggressive futility pruning, but found it decreased the playing strength. Next, I restored late move pruning and left in the aggressive reductions. I found the strength increased, but not quite back to where it had been. I discovered a non-linear relationship between these search techniques. The combined strength of null move reductions, LMR, LMP, and futility pruning is not additive. I experimented with various parameter configurations before deciding to abandon LMP. It took 300 hours of testing to confirm the revised search was equivalently strong to the search with LMP.
This added 20 Elo to the playing strength of MadChess 2.0 Beta.
MadChess 2.0 2306.9 : 1600 (+835,=299,-466), 61.5 % vs. : games ( +, =, -), (%) : Diff, SD, CFS (%) Glass 1.6 : 100 ( 22, 24, 54), 34.0 : -140.1, 15.0, 0.0 RomiChess P3L : 100 ( 20, 20, 60), 30.0 : -133.7, 16.2, 0.0 OliThink 5.3.2 : 100 ( 23, 20, 57), 33.0 : -109.7, 15.5, 0.0 Myrddin 0.87 : 100 ( 31, 21, 48), 41.5 : -75.0, 17.3, 0.0 Sungorus 1.4 : 100 ( 49, 15, 36), 56.5 : -3.1, 8.2, 35.2 FireFly v2.6.0 : 100 ( 47, 31, 22), 62.5 : +89.0, 10.0, 100.0 ZCT 0.3.2450 : 100 ( 42, 28, 30), 56.0 : +94.7, 10.8, 100.0 Beowulf 2.4 : 100 ( 59, 19, 22), 68.5 : +97.1, 13.0, 100.0 Jazz v444 : 100 ( 60, 17, 23), 68.5 : +108.9, 10.0, 100.0 Brainless 1.0 : 100 ( 57, 17, 26), 65.5 : +148.2, 17.8, 100.0 Wing 2.0 : 100 ( 61, 20, 19), 71.0 : +177.5, 10.1, 100.0 BikJump v2.01 : 100 ( 61, 22, 17), 72.0 : +198.9, 10.7, 100.0 Matheus-2.3 : 100 ( 66, 19, 15), 75.5 : +220.5, 10.7, 100.0 Monarch 1.7 : 100 ( 72, 12, 16), 78.0 : +245.5, 11.0, 100.0 BigLion 2.23w : 100 ( 82, 7, 11), 85.5 : +277.1, 9.4, 100.0 Sharper 0.17 : 100 ( 83, 7, 10), 86.5 : +289.8, 10.8, 100.0
Feature | Category | Date | Rev1 | WAC2 | Elo Rating3 | Improvement |
---|---|---|---|---|---|---|
Rook And Queen On 7th Rank | Evaluation | 2015 Aug 20 | 123 | 279 | 2307 | +20 |
Knight Outposts | Evaluation | 2015 May 27 | 83 | 276 | 2287 | +25 |
Tune Reductions And Pruning | Search | 2015 Apr 26 | 74 | 275 | 2262 | +20 |
Bishop Pair | Evaluation | 2015 Apr 15 | 69 | 275 | 2242 | +18 |
Free Passed Pawns | Evaluation | 2015 Mar 10 | 59 | 270 | 2224 | +31 |
Unstoppable Pawns Draws, Material Trades |
Evaluation | 2015 Jan 31 | 52 | 270 | 2193 | +39 |
Late Move Pruning | Search | 2015 Jan 10 | 44 | 273 | 2154 | +39 |
History Heuristic Late Move Reductions |
Search | 2015 Jan 04 | 40 | 275 | 2115 | +50 |
Killer Moves | Search | 2015 Jan 03 | 38 | 275 | 2065 | +61 |
Futility Pruning | Search | 2014 Dec 29 | 37 | 256 | 2004 | +54 |
Null Move Quiescence Recaptures |
Search | 2014 Dec 28 | 34 | 242 | 1950 | +46 |
King Safety | Evaluation | 2014 Dec 24 | 32 | 225 | 1904 | +27 |
Piece Mobility | Evaluation | 2014 Dec 16 | 29 | 225 | 1877 | +64 |
Draw By Repetition Bug | Evaluation | 2014 Dec 10 | 27 | 225 | 1813 | +47 |
Passed Pawns | Evaluation | 2014 Dec 09 | 26 | 225 | 1766 | +72 |
Time Management | Search | 2014 Dec 08 | 25 | 231 | 1694 | +24 |
Delay Move Generation Aspiration Window Bug |
Search | 2014 Dec 02 | 23 | 231 | 1670 | +44 |
MVV / LVA Move Order Draw By Insufficient Material Move List Overflow Bug |
Search | 2014 Dec 01 | 22 | 235 | 1626 | +30 |
Tapered Evaluation MG and EG Piece Location |
Evaluation | 2014 Nov 29 | 21 | 234 | 1596 | +107 |
Alpha / Beta Negamax Aspiration Windows Quiescence, Hash Material, Piece Squares |
Baseline | 2014 Nov 25 | 20 | 236 | 1489 | – |
- Subversion source code revision
- Win At Chess position test, 3 seconds per position
- Bullet chess, 2 min / game + 1 sec / move