Stockfish Testing Queue

Pending - 0 tests 0.0 hrs

None

Active - 0 tests

Finished - 523 tests

16-02-18 II tmm diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 48216 W: 10616 L: 10590 D: 27010
sprt @ 10+0.1 th 1 My take on time management.
16-02-18 II RookOnFile diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 19814 W: 4284 L: 4370 D: 11160
sprt @ 10+0.1 th 1 A little rook tweak.
16-02-18 II nullmove diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 14049 W: 3014 L: 3076 D: 7959
sprt @ 10+0.1 th 1 Null move condition #2
16-02-18 II nullmove diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 5112 W: 1035 L: 1141 D: 2936
sprt @ 10+0.1 th 1 Null move condition #1
09-02-18 II master diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 38230 W: 7786 L: 7804 D: 22640
sprt @ 10+0.1 th 1 Most probably the last try with contempt.
09-02-18 II master diff
LLR: -0.66 (-2.94,2.94) [0.00,4.00]
Total: 5801 W: 1211 L: 1225 D: 3365
sprt @ 10+0.1 th 1 Contempt=12
08-02-18 II master diff
LLR: 1.38 (-2.94,2.94) [-3.00,1.00]
Total: 71353 W: 14541 L: 14599 D: 42213
sprt @ 10+0.1 th 1 I want to check a scaling of the contempt. C=25 failed LTC non-regression. Can it pass STC?
08-02-18 II master diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 25556 W: 5011 L: 5078 D: 15467
sprt @ 10+0.1 th 1 OK, now when we have a default contempt, I'll try few values with [0,4] bounds, starting with the most logical one: C=0.
07-02-18 II assorted diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 31175 W: 6982 L: 7022 D: 17171
sprt @ 10+0.1 th 1 Take 3.
06-02-18 II assorted diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 21528 W: 4748 L: 4827 D: 11953
sprt @ 10+0.1 th 1 I want to take advantage of recent Elo tests by Fauzi Akram for some manual tuning, take 2.
05-02-18 II assorted diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 30252 W: 6731 L: 6775 D: 16746
sprt @ 10+0.1 th 1 I want to take advantage of recent Elo tests by Fauzi Akram for some manual tuning, take 1.
29-01-18 II pawn_tweak diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 34599 W: 7598 L: 7626 D: 19375
sprt @ 10+0.1 th 1 Tuned values, take 2.
24-01-18 II tune_lowc diff
57561/60000 iterations
120000/120000 games played
120000 @ 10+0.1 th 1 Continue tuning...
27-01-18 II grf diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 1912 W: 376 L: 499 D: 1037
sprt @ 10+0.1 th 1 General reduction formula, take 2.
27-01-18 II grf diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 1927 W: 362 L: 484 D: 1081
sprt @ 10+0.1 th 1 General reduction formula, take 1.
24-01-18 II pawn_tweak diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 29710 W: 6556 L: 6603 D: 16551
sprt @ 10+0.1 th 1 The tuning was stopped, but let's try it.
22-01-18 II tune_lowc diff
38244/60000 iterations
81431/120000 games played
120000 @ 10+0.1 th 1 Last few tries to use SPSA effectively, now with low ck values.
20-01-18 II nms diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 61222 W: 11078 L: 11026 D: 39118
sprt @ 10+0.1 th 1 Null move search - last try so far.
16-01-18 II nms diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 28393 W: 5042 L: 5048 D: 18303
sprt @ 10+0.1 th 1 Some of tuned values - corrected version.
16-01-18 II nms diff
LLR: -0.87 (-2.94,2.94) [0.00,5.00]
Total: 1195 W: 200 L: 232 D: 763
sprt @ 10+0.1 th 1 Some of tuned values.
15-01-18 II nms diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 8092 W: 1424 L: 1515 D: 5153
sprt @ 10+0.1 th 1 4th test in the series on null move search
14-01-18 II nms diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 5760 W: 1006 L: 1107 D: 3647
sprt @ 10+0.1 th 1 3rd test in the series on null move search
13-01-18 II master diff
ELO: 300.85 +-3.5 (95%) LOS: 100.0%
Total: 40000 W: 29191 L: 1219 D: 9590
40000 @ 10+0.1 th 1 Master Contempt 7 vs Stockfish 5
13-01-18 II master diff
ELO: 293.06 +-3.4 (95%) LOS: 100.0%
Total: 40000 W: 28790 L: 1284 D: 9926
40000 @ 10+0.1 th 1 Master vs Stockfish 5
13-01-18 II tmm diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 25218 W: 4355 L: 4553 D: 16310
sprt @ 10+0.1 th 1 The second and the last try.
13-01-18 II master diff
Pending...
sprt @ 10+0.1 th 1 Master Contempt 7 vs Stockfish 5
13-01-18 II tmm diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 11756 W: 2021 L: 2197 D: 7538
sprt @ 10+0.1 th 1 Try a small simplification, a similar to one tried by lucasart for the previous version of time management.
12-01-18 II noadj diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 21370 W: 4412 L: 4493 D: 12465
sprt @ 10+0.1 th 1 As contempt tests are going fine, I decided to try Contempt 20 with SPRT[0,4] bounds.
12-01-18 II nms diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 20675 W: 3707 L: 3793 D: 13175
sprt @ 10+0.1 th 1 2nd test in the series on null move search
11-01-18 II nms diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 77090 W: 14044 L: 13937 D: 49109
sprt @ 10+0.1 th 1 1st test in the series on null move search
09-01-18 II noadj diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 13045 W: 1834 L: 1703 D: 9508
sprt @ 60+0.6 th 1 LTC: Contempt=7 test without adjudication rules. Most probably the last test before pull request.
09-01-18 II noadj diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 50665 W: 9813 L: 9745 D: 31107
sprt @ 10+0.1 th 1 Repeat Contempt=7 test without adjudication rules.
07-01-18 II noadj diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 38185 W: 5507 L: 5548 D: 27130
sprt @ 60+0.6 th 1 Approximation of recent Elo tests by xoto10 gives Contempt=12 as the best value on STC. As framework is almost empty at the moment, and as Contempt=10 was yellow on STC, try LTC for Contempt=12 with low throughput.
07-01-18 II noadj diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 66299 W: 13110 L: 13028 D: 40161
sprt @ 10+0.1 th 1 Contempt=10 with SPRT[0,4] bounds.
07-01-18 II noadj diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 21416 W: 2928 L: 2808 D: 15680
sprt @ 60+0.6 th 1 LTC: I'm trying to prove that small positive contempt is good also in self-play. Test without adjudication rules, starting from Contempt=4.
06-01-18 II noadj diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 63638 W: 12048 L: 12002 D: 39588
sprt @ 10+0.1 th 1 I'm trying to prove that small positive contempt is good also in self-play. Test without adjudication rules, starting from Contempt=4.
05-01-18 II contempt diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 111553 W: 20712 L: 21065 D: 69776
sprt @ 10+0.1 th 1 As there is some feeling among people (and some recent tests with 20K games indicate that) that small positive contempt do not lose Elo (for no clear reason, of course) - test it for no regression
30-12-17 II tune_nmp diff
174740/250000 iterations
380021/500000 games played
500000 @ 10+0.1 th 1 Big tuning of important concept of null move search. Half-throughput and I'll stop earlier if nothing important happens.
30-12-17 II oldtmm diff
LLR: 3.44 (-2.94,2.94) [-3.00,1.00]
Total: 31611 W: 3958 L: 3827 D: 23826
sprt @ 60+0.6 th 1 LTC for time management revert
30-12-17 II oldtmm diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 14060 W: 2562 L: 2430 D: 9068
sprt @ 10+0.1 th 1 STC for time management revert
29-12-17 II master diff
ELO: -0.90 +-4.6 (95%) LOS: 35.0%
Total: 5000 W: 559 L: 572 D: 3869
5000 @ 36+0.05 th 7 TCEC ready test: after some discussion under issue #1272 and after some local tests I suspect that SMP could be a problem for new time management. This time control is TCEC 11 Superfinal divided by 200, and it is also close to sudden death case, which is under suspicion by some users.
09-12-17 II contem diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 109961 W: 14232 L: 14095 D: 81634
sprt @ 60+0.6 th 1 LTC: Take 2: Larger change
12-12-17 II contem diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 39958 W: 7273 L: 7293 D: 25392
sprt @ 10+0.1 th 1 Last try on this idea: (Tempo, Contempt) = (24, 2).
12-12-17 II evalcon diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 6651 W: 1307 L: 1405 D: 3939
sprt @ 10+0.1 th 1 Evaluation based contempt - take 2.
11-12-17 II evalcon diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 5854 W: 1084 L: 1185 D: 3585
sprt @ 10+0.1 th 1 Evaluation based contempt - take 1.
09-12-17 II tempo diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 84768 W: 15355 L: 15223 D: 54190
sprt @ 10+0.1 th 1 Is Stockfish ready for higher tempo? (an alternative / check for running LTC test)
10-12-17 II stempo diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 18955 W: 2367 L: 2424 D: 14164
sprt @ 60+0.6 th 1 LTC: Define Tempo as Score.
10-12-17 II stempo diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 9323 W: 1749 L: 1585 D: 5989
sprt @ 10+0.1 th 1 Define Tempo as Score.
09-12-17 II contem diff
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 25041 W: 4710 L: 4467 D: 15864
sprt @ 10+0.1 th 1 Take 2: Larger change
09-12-17 II contem diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 23666 W: 4256 L: 4332 D: 15078
sprt @ 10+0.1 th 1 Small increase for contempt and tempo, based on tuning graphs.