Stockfish Testing Queue

Finished - 27247 tests

30-10-15 SC quadratic_initiative_tu diff
46131/100000 iterations
137678/200000 games played
200000 @ 10+0.1 th 1 Fix base branch. Not my day, in any case.
30-10-15 jo easymove_tweak diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 129795 W: 24983 L: 24679 D: 80133
sprt @ 10+0.1 th 1 Take 3.
30-10-15 pe tm diff
LLR: -4.41 (-2.94,2.94) [0.00,4.00]
Total: 72115 W: 13874 L: 13853 D: 44388
sprt @ 10+0.1 th 1 Try simple decrease of time per move
30-10-15 Vo HPtweak1b diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 9313 W: 1730 L: 1816 D: 5767
sprt @ 10+0.1 th 1 Try to improve History-pruning...
30-10-15 jo stop_iteration diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 23645 W: 4495 L: 4569 D: 14581
sprt @ 10+0.1 th 1 Before eventually running a combo patch, let's first see how this is doing. Take 2, already stop at 80%.
30-10-15 Vo SBF3c diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 23058 W: 4457 L: 4533 D: 14068
sprt @ 10+0.1 th 1 Same as 3b. But remove FL bonus. Curious to see what happens.
30-10-15 jo stop_iteration diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 97063 W: 18619 L: 18315 D: 60129
sprt @ 10+0.1 th 1 Take 2. Don't start a new iteration, if we already spent more than 90% of our available time.
30-10-15 jo easymove_tweak^ diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 47535 W: 9074 L: 9064 D: 29397
sprt @ 10+0.1 th 1 Take 2.
30-10-15 jo easymove_tweak diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 40574 W: 7723 L: 7737 D: 25114
sprt @ 10+0.1 th 1 Allow consecutive easy moves, but spend more time.
30-10-15 lb no_timer^ diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 2955 W: 733 L: 857 D: 1365
sprt @ 1.2+0.02 th 15 Each thread polls every 4k nodes: yes that means we poll N times more with N threads, but I bet the slowdown is not mesurable. And the more threads the more conservative we need to be, to protect ourselves against OS scheduler playing tricks on us. To test this, we need a very short tc, with many threads.
30-10-15 aj SBF3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 13788 W: 1972 L: 2043 D: 9773
sprt @ 60+0.4 th 1 LTC test against master: Rescheduling VoyagerOne's test that passed STC
30-10-15 SC quadratic_initiative_tu diff
10/50000 iterations
16/100000 games played
100000 @ 10+0.1 th 1 Bugfix, thanks to Rocky640
30-10-15 Ro WeakPinned diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 27331 W: 5260 L: 5265 D: 16806
sprt @ 10+0.1 th 1 a one time penalty if we have a pinned piece which is not pawn supported
28-10-15 SC opposite_bishops_simple diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 29414 W: 5476 L: 5686 D: 18252
sprt @ 10+0.1 th 1 Variation on Stephane's idea of removing ad-hoc logic for opposite bishops. Take 2.
28-10-15 SC opposite_bishops_simple diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 25206 W: 4700 L: 4903 D: 15603
sprt @ 10+0.1 th 1 Variation on Stephane's idea of removing ad-hoc logic for opposite bishops.
30-10-15 Vo SBF3 diff
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 109459 W: 21288 L: 20738 D: 67433
sprt @ 10+0.1 th 1 Try: d*d+2*d-2. Test against passed SBF2.
30-10-15 Se cutoff_time diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 16881 W: 3309 L: 3360 D: 10212
sprt @ 10+0.1 th 1 Suggested by user IIvec. Same as the previous two yellow tests, but with divisor 5.
30-10-15 Se cutoff_time_div8 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 32054 W: 6189 L: 6173 D: 19692
sprt @ 10+0.1 th 1 Same as previous yellow "cutoff_time", but with divisor 8 instead of 10. Let's see if it is better.
29-10-15 SC quadratic_initiative_tu diff
4002/50000 iterations
8403/100000 games played
100000 @ 10+0.1 th 1 Try to tune a generalization of evaluate_initiative: every contribution is now a score and also quadratic terms are considered.
27-10-15 Ro TempoVerification diff
LLR: -3.57 (-2.94,2.94) [0.00,4.00]
Total: 31234 W: 4511 L: 4603 D: 22120
sprt @ 60+0.4 th 1 Is tempo=17 better at LTC ?
29-10-15 Se cutoff_time diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 32866 W: 6294 L: 6275 D: 20297
sprt @ 10+0.1 th 1 This is my suggestion to prevent losses on time. It uses a formula for cutoffTime so that check_time() will stop searching before it is too late even when timer thread is lagging. See also my posts in this thread: https://groups.google.com/forum/#!topic/fishcooking/kPNocZQkW-4
29-10-15 Vo SBF2 diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 21225 W: 3196 L: 3000 D: 15029
sprt @ 60+0.4 th 1 LTC: Formula 2: d*d+d-1
29-10-15 Fi play_on_increment2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 22337 W: 4279 L: 4306 D: 13752
sprt @ 10+0.1 th 1 Improved time management based on Leonid's patch.
29-10-15 Vo US diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 19947 W: 3791 L: 3829 D: 12327
sprt @ 10+0.1 th 1 Update Stat Tweak: Note: Depth is too low for bench to change.
29-10-15 SC opposite_bishops_simple diff
LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 7503 W: 1341 L: 1513 D: 4649
sprt @ 10+0.1 th 1 A last go with tuned values.
28-10-15 Vo Improving diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 114820 W: 22504 L: 22114 D: 70202
sprt @ 10+0.1 th 1 Increase stats bonus if improving.
28-10-15 SC oppbis_simple_tuning diff
58047/50000 iterations
100000/100000 games played
100000 @ 5+0.1 th 1 Variation on Stephane's idea of removing ad-hoc logic for opposite bishops. Try to get the right coefficients by tuning at fast time control.
29-10-15 Vo SBF2 diff
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 5916 W: 1199 L: 1046 D: 3671
sprt @ 10+0.1 th 1 Formula 2: d*d+d-1
29-10-15 mc no_timer diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 36453 W: 7601 L: 7508 D: 21344
sprt @ 2+0.05 th 7 Retire timer. New approach. Almost zero increment.
29-10-15 Vo SBF diff
LLR: -2.18 (-2.94,2.94) [0.00,5.00]
Total: 6856 W: 1263 L: 1326 D: 4267
sprt @ 10+0.1 th 1 Use move count formula as a way to score stats. The slope is shallower than x^2 and we will be able to score stats at higher depths.
29-10-15 mc no_timer diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 26511 W: 4465 L: 4353 D: 17693
sprt @ 10+0.1 th 7 Retire timer. New approach. Normal time.
29-10-15 tv no_timer diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 20433 W: 4246 L: 4123 D: 12064
sprt @ 2+0.05 th 7 Variant with uniform distribution
29-10-15 mc no_timer diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 19586 W: 4108 L: 3983 D: 11495
sprt @ 2+0.05 th 7 Retire timer. New approach without no now(). Almost zero increment.
29-10-15 jo stop_iteration diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 12250 W: 2296 L: 2369 D: 7585
sprt @ 10+0.1 th 1 Does it help to not start a new iteration if we already spent more than 90% of our maximum time?
29-10-15 II spare_inc diff
ELO: -1.23 +-2.8 (95%) LOS: 19.3%
Total: 28149 W: 6598 L: 6698 D: 14853
30000 @ 2+0.05 th 1 Stressing this patch. I think this patch can also be helpful in our testings, by lowering the number of time losses. This patch is doing well and will always leave (7500 - Tick) ms on Martin's machine.
28-10-15 II spare_inc diff
LLR: 1.19 (-2.94,2.94) [-3.00,1.00]
Total: 77546 W: 11348 L: 11399 D: 54799
sprt @ 60+0.4 th 1 This patch is doing well and will always leave (7500 - Tick) ms on Martin's machine.
29-10-15 lb no_timer diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 1514 W: 499 L: 327 D: 688
sprt @ 1+0.02 th 7 Retire timer. New approach without no now(). Almost zero increment. => shortest possible tc that should work in fishtest. note that fishtest does not purge time losses anymore.
29-10-15 pe play_on_increment diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 14495 W: 2772 L: 2834 D: 8889
sprt @ 10+0.1 th 1 Allow using 6% more time for most of the game because of better increment play
29-10-15 pe play_on_increment diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 15105 W: 2917 L: 2976 D: 9212
sprt @ 10+0.1 th 1 Allow using more time for most of the game because of better increment play
29-10-15 mc no_timer diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 7073 W: 1761 L: 1607 D: 3705
sprt @ 2+0.05 th 1 Retire timer. New approach. Almost zero increment.
29-10-15 pe play_on_increment diff
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 3245 W: 745 L: 600 D: 1900
sprt @ 0.25+0.25 th 1 Check that effect is still there if condition moved to another place
28-10-15 SC moreTuned diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 54730 W: 10623 L: 10585 D: 33522
sprt @ 10+0.1 th 1 Results from SPSA tuning session. Keep symmetrized pawn values and remove history tuned values, according to mixed suggestions by VOne and Rocky. Take 4.
29-10-15 lb polling diff
ELO: -36.07 +-3.9 (95%) LOS: 0.0%
Total: 17265 W: 3891 L: 5677 D: 7697
30000 @ 1.2+0.02 th 7 see patch comments
28-10-15 jh time_cleanup diff
LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 37830 W: 7193 L: 7420 D: 23217
sprt @ 10+0.1 th 1 Tweak a couple of parameters.
28-10-15 jo cap_available diff
LLR: -2.94 (-2.94,2.94) [0.00,4.00]
Total: 8501 W: 1579 L: 1707 D: 5215
sprt @ 10+0.1 th 1 Don't allow available to become greater than maximum. Supposed to only affect time decision for easy moves.
28-10-15 n_ ComboTune diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 16824 W: 3247 L: 3299 D: 10278
sprt @ 10+0.1 th 1 A combo patch with all tuningpatches finishing yellow since the last … A combo patch with all tuning patches finishing yellow since the last succesful combo pach. Testing with [0,5].
28-10-15 sg passed_pawns diff
LLR: -3.61 (-2.94,2.94) [0.00,5.00]
Total: 34580 W: 6625 L: 6626 D: 21329
sprt @ 10+0.1 th 1 Rework safe squares bonus. Take 2. For the master following values would be equivalent SafeSquares[] = {18, 0, 8, 8, 8, 8, 8}. Use this time only slightly different values. Remark: the two last values are currently never used because there can only occur for pawns on rank 2 or 3 and in this case the multiplicator rr is zero.
28-10-15 sn holes diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 23746 W: 4581 L: 4602 D: 14563
sprt @ 10+0.1 th 1 Take 2 (lower malus)
28-10-15 mc no_timer diff
LLR: -2.61 (-2.94,2.94) [-3.00,1.00]
Total: 9515 W: 1854 L: 2014 D: 5647
sprt @ 2+0.05 th 7 Retire timer. Final version. Almost zero increment.
28-10-15 mc no_timer diff
LLR: -0.18 (-2.94,2.94) [-3.00,1.00]
Total: 3266 W: 550 L: 565 D: 2151
sprt @ 10+0.1 th 7 Retire timer. Final version.