Stockfish Testing Queue

Finished - 1777 tests

18-03-28 pro ps_pawnsupport diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 29111 W: 6050 L: 6042 D: 17019
sprt @ 10+0.1 th 1 Can pawns behind us can help pass pawns?
18-03-28 pro ps_sorting_tune diff
16738/20000 iterations
33714/40000 games played
40000 @ 20+0.2 th 1 values moved in first run.. continue tuning. Prio: -1.
18-03-28 pro ps_sorting_tune diff
9563/10000 iterations
19279/20000 games played
20000 @ 20+0.2 th 1 try to tune sorting limits.
18-03-28 pro ps_shelter11 diff
LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 99399 W: 20393 L: 20743 D: 58263
sprt @ 10+0.1 th 1 for information only, test for non-regression of simplified pawn shelter array.
18-03-28 pro ps_shelter11_tune diff
8640/10000 iterations
17384/20000 games played
20000 @ 20+0.2 th 1 simplified array still converging.
18-03-28 pro ps_shelter11_tune diff
8561/10000 iterations
17233/20000 games played
20000 @ 20+0.2 th 1 simplified shelter: Test tune a heuristic array for comparison.
18-03-28 pro ps_shelter11_tune diff
9928/10000 iterations
20000/20000 games played
20000 @ 20+0.2 th 1 OK, now tune the whole simplified 4x4 array.
18-03-28 pro ps_shelter11_tune diff
2585/10000 iterations
5231/20000 games played
20000 @ 20+0.2 th 1 tune safety base before tuning shelter array.
18-03-28 pro ps_shelter11_tune diff
9272/10000 iterations
18696/20000 games played
20000 @ 10+0.1 th 1 tuning run #2 (updated values). tune simplified pawn shelter.
18-03-27 pro ps_shelter11_tune diff
8930/10000 iterations
17974/20000 games played
20000 @ 20+0.2 th 1 tune simplified pawn shelter.
18-03-27 pro ps_shelter11 diff
LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 51113 W: 10397 L: 10655 D: 30061
sprt @ 10+0.1 th 1 simplify away deeper pawns in ShelterWeakness (now ShelterStrength).
18-03-27 pro ps_shelter9 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 39566 W: 8160 L: 8172 D: 23234
sprt @ 10+0.1 th 1 Testing relevance of deeper pawns on ShelterWeakness. Probably won't pass, but is good information.
18-03-27 pro ps_shelter8 diff
LLR: -0.30 (-2.94,2.94) [-3.00,1.00]
Total: 423 W: 71 L: 88 D: 264
sprt @ 10+0.1 th 1 Try#2: try to simply kingOnFile factor in shelterweakness.
18-03-26 pro ps_shelter3 diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 100690 W: 15373 L: 15372 D: 69945
sprt @ 60+0.6 th 1 LTC: try to encode shelter/storm masks into Danger score.
18-03-26 pro ps_shelter3 diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 36597 W: 7522 L: 7429 D: 21646
sprt @ 10+0.1 th 1 try to encode shelter/storm masks into Danger score.
18-03-26 pro ps_shelter3 diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 23402 W: 4890 L: 4772 D: 13740
sprt @ 10+0.1 th 1 test value of this special rule.
18-03-26 pro ps_shelter2 diff
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 97105 W: 20035 L: 20057 D: 57013
sprt @ 10+0.1 th 1 simplify ShelterWeakness
18-03-26 pro ps_shelter2 diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 22763 W: 4645 L: 4850 D: 13268
sprt @ 10+0.1 th 1 v2: simplify shelter weakness.
18-03-25 pro ps_tune_shelter diff
846/10000 iterations
1758/21000 games played
21000 @ 10+0.1 th 1 Still converging. Run#2. tune shelter weakness equation.
18-03-24 pro ps_tune_shelter diff
9013/10000 iterations
20000/20000 games played
20000 @ 10+0.1 th 1 tune interpolated shelter equation (more movement).
18-03-24 pro ps_tune_shelter diff
208/10000 iterations
1426/20000 games played
20000 @ 20+0.2 th 1 tune interpolated shelter equation.
18-03-24 pro ps_tune_shelter diff
56/10000 iterations
333/20000 games played
20000 @ 20+.2 th 1 first tuning. Delete if this breaks something.
18-03-24 pro ps_backward3 diff
LLR: -0.16 (-2.94,2.94) [0.00,5.00]
Total: 673 W: 98 L: 102 D: 473
sprt @ 60+0.6 th 1 Marco wants LTC for this non-functional simplification. Don't determine "backwardness" until it's needed.
18-03-23 pro ps_backward3 diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 27015 W: 5610 L: 5499 D: 15906
sprt @ 10+0.1 th 1 simplify backward pawns (don't test unless we need to).
18-03-23 pro ps_backward3 diff
LLR: -0.52 (-2.94,2.94) [-3.00,1.00]
Total: 1746 W: 250 L: 279 D: 1217
sprt @ 60+0.6 th 1 LTC: simplify backward pawns. I'm certain STC will pass. Submitting early.
18-03-23 pro ps_passed2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 7508 W: 1492 L: 1586 D: 4430
sprt @ 10+0.1 th 1 hard/soft passed pawns with bonus for hard passed duos. S(10,10);
18-03-21 pro ps_scaledlazy4 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 82121 W: 10013 L: 9878 D: 62230
sprt @ 240+2.4 th 1 on empty framework, just see if there is any elo here. Throughput 50, Prio -10.
18-03-23 pro ps_passedneighbours diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 16356 W: 3362 L: 3414 D: 9580
sprt @ 10+0.1 th 1 last try. this is dying quickly. lol. S(20,20)
18-03-23 pro ps_passedneighbours diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 26360 W: 5395 L: 5401 D: 15564
sprt @ 10+0.1 th 1 extra bonus for passed pawns that are also neighbours. S(10,10)
18-03-23 pro ps_passedneighbours diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 6822 W: 1361 L: 1458 D: 4003
sprt @ 10+0.1 th 1 bonus for passed neighbours S(4,4).
18-03-20 pro ps_lazyking diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 11291 W: 2277 L: 2353 D: 6661
sprt @ 10+0.1 th 1 try#3: mg > 0, no eg score.
18-03-20 pro ps_lazyking diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 6538 W: 1284 L: 1382 D: 3872
sprt @ 10+0.1 th 1 don't do lazy eval if king is not super safe. 30,-17
18-03-20 pro ps_lazyking diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 6125 W: 1216 L: 1317 D: 3592
sprt @ 10+0.1 th 1 don't do lazy eval if king low king safety. 100,-17
18-03-18 pro ps_targetking diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 9082 W: 1818 L: 1905 D: 5359
sprt @ 10+0.1 th 1 S(10,0) bonus or pieces that would attack king if all other pieces were gone.
18-03-18 pro ps_targetking diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 16960 W: 3378 L: 3428 D: 10154
sprt @ 10+0.1 th 1 bonus for pieces that attack king ring with all other pieces removed. S(5,0)
18-03-18 pro ps_targetking diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 9406 W: 1964 L: 2049 D: 5393
sprt @ 10+0.1 th 1 targetking with other pieces removed S(0,5)
18-03-17 pro ps_weakpawns2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 16755 W: 3385 L: 3436 D: 9934
sprt @ 10+0.1 th 1 bonus S(2,8) for attacking weak pawns.
18-03-17 pro ps_weakpawns2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 7608 W: 1568 L: 1662 D: 4378
sprt @ 10+0.1 th 1 bonus S(7,7) for attacking isolated or backward pawns.
18-03-17 pro ps_scaledlazy2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 18623 W: 2754 L: 2807 D: 13062
sprt @ 60+0.6 th 1 speculative LTC, low priority, low throughput, use depth to scale lazyeval, 1400+50 per depth.
18-03-17 pro ps_rams diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 13009 W: 2680 L: 2748 D: 7581
sprt @ 10+0.1 th 1 introduce pawn rams: bonus for enemy pawns on the same color as our bishop.
18-03-17 pro ps_rams diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 26238 W: 5343 L: 5349 D: 15546
sprt @ 10+0.1 th 1 try#4: introduce rams: S(3,3)
18-03-17 pro ps_scaledlazy3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 7791 W: 1146 L: 1238 D: 5407
sprt @ 60+0.6 th 1 speculative LTC, prio: -1, throughput 200, try to scale lazyEval using rootDepth.
18-03-17 pro ps_rams diff
LLR: -3.33 (-2.94,2.94) [0.00,5.00]
Total: 33805 W: 6858 L: 6846 D: 20101
sprt @ 10+0.1 th 1 try#3: introduce rams. less penalty S(5,5).
18-03-17 pro ps_rams diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 16384 W: 3419 L: 3471 D: 9494
sprt @ 10+0.1 th 1 introduce rams. increase penalty to S(15,15).
18-03-16 pro ps_rams diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 27434 W: 5723 L: 5722 D: 15989
sprt @ 10+0.1 th 1 sry, bugfix again. :( introduce pawn rams. double penalty for rams on the same color as a bishop.
18-03-16 pro ps_rams diff
LLR: 0.20 (-2.94,2.94) [0.00,5.00]
Total: 1399 W: 281 L: 266 D: 852
sprt @ 10+0.1 th 1 introduce pawn rams. double penalty for rams on the same color as a bishop.
18-03-16 pro ps_scaledlazy3 diff
LLR: -0.30 (-2.94,2.94) [0.00,5.00]
Total: 1887 W: 298 L: 303 D: 1286
sprt @ 60+0.6 th 1 speculative LTC. prio: -1. low throughput, use rootDepth to scale lazyEval. 1400+200 for depth > 10.
18-03-14 pro ps_scaledlazy diff
LLR: -1.52 (-2.94,2.94) [0.00,5.00]
Total: 71962 W: 8867 L: 8705 D: 54390
sprt @ 240+2.4 th 1 now test for elo gain at VLTC. super low throughput (50).
18-03-14 pro ps_pbsimple diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 93957 W: 19252 L: 19062 D: 55643
sprt @ 10+0.1 th 1 non-functional: If there is only one move, just return it.
18-03-13 pro ps_scaledlazy diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 35520 W: 5469 L: 5370 D: 24681
sprt @ 60+0.6 th 1 trying to scale away lazythreshold at long games. Check if this is a non-regression at 60+.6 before testing VLTC. low throughput.