Stockfish Testing Queue

Finished - 49050 tests

16-08-03 tvi futil diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 16052 W: 2960 L: 3062 D: 10030
sprt @ 10+0.1 th 1 Take 4: return eval - FM() / 2
16-08-02 jos probcut diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 42139 W: 5666 L: 5701 D: 30772
sprt @ 60+0.6 th 1 Revisit ProbCut. LTC
16-08-03 Elb rook_space diff
LLR: -0.88 (-2.94,2.94) [0.00,5.00]
Total: 3380 W: 456 L: 480 D: 2444
sprt @ 60+0.6 th 1 Take 2 (LTC)
16-08-03 IIv futility diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 9284 W: 1706 L: 1792 D: 5786
sprt @ 10+0.1 th 1 Futility tweak.
16-08-02 aji checks diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 49782 W: 9353 L: 9262 D: 31167
sprt @ 10+0.1 th 1 Always penalize safe/other checks: STC
16-08-03 Elb rook_space diff
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 8858 W: 1811 L: 1643 D: 5404
sprt @ 10+0.1 th 1 Take 2
16-08-03 pec fail_high3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 11410 W: 1923 L: 2000 D: 7487
sprt @ 5+0.05 th 7 Try at 7threads this also. passed 1th stc & ltc, and 3threads stc
16-08-03 jos null_end2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 19245 W: 3575 L: 3617 D: 12053
sprt @ 10+0.1 th 1 Limit reduction in verification search. Take 2.
16-08-03 Elb rook_space diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 11482 W: 2127 L: 2203 D: 7152
sprt @ 10+0.1 th 1 Experiment with rook space
16-08-02 tvi futil diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 23648 W: 3159 L: 3198 D: 17291
sprt @ 60+0.6 th 1 LTC: Rebased, take 2
16-08-02 aji big_threats_tuned_fix diff
LLR: -3.52 (-2.94,2.94) [0.00,5.00]
Total: 9959 W: 1775 L: 1882 D: 6302
sprt @ 10+0.1 th 1 big threats tuned values: STC(fix bug)
16-08-02 jos null_end2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 18376 W: 3430 L: 3476 D: 11470
sprt @ 10+0.1 th 1 Decrease null-reduction when low on non-pawn-material. Take 1.
16-08-02 jos probcut diff
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 10400 W: 2076 L: 1881 D: 6443
sprt @ 10+0.1 th 1 Revisit ProbCut.
16-08-02 Voy rpd diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 13454 W: 2380 L: 2560 D: 8514
sprt @ 10+0.1 th 1 Remove predicted depth
16-08-02 tvi futil diff
LLR: -0.16 (-2.94,2.94) [0.00,5.00]
Total: 429 W: 79 L: 84 D: 266
sprt @ 10+0.1 th 1 LTC: Rebased, take 2
16-08-02 tvi futil diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 21310 W: 4167 L: 3946 D: 13197
sprt @ 10+0.1 th 1 Take 2, fixed logic
16-08-02 pb0 score_capture_on_hangin diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 8065 W: 1450 L: 1541 D: 5074
sprt @ 10+0.1 th 1 Score bonus for captures on hanging pieces, take 2
16-08-02 aji big_threats_tuned diff
LLR: -0.80 (-2.94,2.94) [0.00,5.00]
Total: 2812 W: 511 L: 533 D: 1768
sprt @ 10+0.1 th 1 Check if tuned values are better : STC
16-08-02 pb0 score_capture_on_hangin diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 7588 W: 1373 L: 1466 D: 4749
sprt @ 10+0.1 th 1 Score bonus for captures on hanging pieces
16-08-01 aji threat_tune diff
28934/30000 iterations
59307/60000 games played
60000 @ 10+0.1 th 1 Tune big threats
16-08-02 tvi futil diff
LLR: -0.63 (-2.94,2.94) [0.00,5.00]
Total: 4966 W: 962 L: 967 D: 3037
sprt @ 10+0.1 th 1 Take 2
16-08-02 Elb pawn_threat diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 4370 W: 787 L: 895 D: 2688
sprt @ 10+0.1 th 1 Pawn under threat
16-08-02 tvi futil diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 22042 W: 4201 L: 4230 D: 13611
sprt @ 10+0.1 th 1 Try a variant of VoyagerOne's futility patch
16-08-01 Voy fpv diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 29331 W: 4019 L: 4038 D: 21274
sprt @ 60+0.6 th 1 I would like to see how this yellow patch fares at LTC... this will give me some ideas on how futility relates with scaling.(low throughput)
16-07-31 Fis easySMP2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 25847 W: 4284 L: 4304 D: 17259
sprt @ 5+0.05 th 7 A stricter version of http://tests.stockfishchess.org/tests/view/5657b8d70ebc5902cdc08546
16-08-01 jos futility_margin2 diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 49488 W: 9354 L: 9338 D: 30796
sprt @ 10+0.1 th 1 Let's see how this tweak works at parent node alone.
16-08-01 Voy fpv3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 18617 W: 3510 L: 3554 D: 11553
sprt @ 10+0.1 th 1 Take 3...
16-08-01 Voy fpv diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 52823 W: 9843 L: 9740 D: 33240
sprt @ 10+0.1 th 1 Fixed logic...
16-08-01 SC double_reductions diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 16273 W: 3050 L: 3105 D: 10118
sprt @ 10+0.1 th 1 Initialize reductions using see
16-08-01 aji big_threats diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 12863 W: 2365 L: 2435 D: 8063
sprt @ 10+0.1 th 1 Some tweaks to handle big threats better. Will probably require tuning. But try with manual values to begin with: STC (fixed bench)
16-08-01 aji big_threats3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13438 W: 2532 L: 2599 D: 8307
sprt @ 10+0.1 th 1 Take 3: STC(user lower values for threat tempo)
16-08-01 aji big_threats2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13352 W: 2546 L: 2613 D: 8193
sprt @ 10+0.1 th 1 Take 2
16-08-01 Voy fpv diff
LLR: -1.89 (-2.94,2.94) [0.00,5.00]
Total: 14038 W: 2662 L: 2681 D: 8695
sprt @ 10+0.1 th 1 Take 2: Try < 5.
16-08-01 pec fail_high4 diff
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 58057 W: 11029 L: 10649 D: 36379
sprt @ 5+0.05 th 3 fix bug in previous patch.
16-08-01 Voy fpv diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 49022 W: 9397 L: 9306 D: 30319
sprt @ 10+0.1 th 1 Use depth < 6 for pv nodes.
16-08-01 pec fail_high3a diff
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 9130 W: 1785 L: 1619 D: 5726
sprt @ 5+0.05 th 3 Should lower depth threads resolve fail high on best move change?
16-08-01 Elb opp_bishop diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 11991 W: 2264 L: 2338 D: 7389
sprt @ 10+0.1 th 1 Take 3
16-07-27 aji smp_thread_skip diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 8899 W: 1183 L: 1272 D: 6444
sprt @ 10+0.1 th 15 This test has run at 23 threads for two weeks! It is unlikely to finish in the near future. Run at 15 threads since there are more such machines and it can realistically complete
16-07-31 Roc ThreatByRook diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 47356 W: 9060 L: 8977 D: 29319
sprt @ 10+0.1 th 1 Tweak to threats by rook. Exclude attacks on pieces defended twice. Bench fixed.
16-07-31 SC pawnSpanInitiative diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 35235 W: 6841 L: 6810 D: 21584
sprt @ 10+0.1 th 1 Further attempt: compensate by reducing pawn weight. (I know I am pushing this quite a lot, but I am somehow encouraged by all patches failing yellow) and (If one finally passes I will be careful in opening a PR).
16-07-31 pec fail_high4 diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 8051 W: 1604 L: 1442 D: 5005
sprt @ 5+0.05 th 3 More elaborate way of determining current best in smp case
16-07-31 SC double_reductions diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 30672 W: 5842 L: 5833 D: 18997
sprt @ 10+0.1 th 1 Values tuned at average LTC.
16-07-31 tvi search diff
LLR: -2.94 (-2.94,2.94) [0.00,4.00]
Total: 23300 W: 4442 L: 4517 D: 14341
sprt @ 10+0.1 th 1 Last shot at this, qsearch futility at 178
16-07-31 lbr outpost diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 16690 W: 3189 L: 3241 D: 10260
sprt @ 10+0.1 th 1 take 2
16-07-31 Voy fmTweak diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 21446 W: 4058 L: 4140 D: 13248
sprt @ 10+0.1 th 1 Use 140 factor...
16-07-30 SC double_reductions_tune diff
26428/30000 iterations
53356/60000 games played
60000 @ 30+0.3 th 1 Try to exploit double valued reductions. Reschedule as not converged As suggested by ElbertoOne, retune at average tc, starting from optimal values until now. Try out randomized rounding. Low throughput.
16-07-31 tvi fbtune diff
7127/10000 iterations
14445/20000 games played
20000 @ 20+0.2 th 1 Try to tune futilityBase
16-07-31 lbr outpost^ diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 16517 W: 3177 L: 3230 D: 10110
sprt @ 10+0.1 th 1 double outpost bonus when opponent has no knight
16-07-31 pec fail_high3 diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 19443 W: 3776 L: 3564 D: 12103
sprt @ 5+0.05 th 3 Another attempt to extend one thread stc and ltc results to smp.
16-07-31 jos lmrt3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 6483 W: 1204 L: 1302 D: 3977
sprt @ 10+0.1 th 1 LMR tweak. Decrease reduction if we are still close to the root. (wrong bench copied)