Nonlinear2 and Joost VandeVondele
4975b2bc6f
Increase prior countermove bonus if TT move
...
Passed STC:
https://tests.stockfishchess.org/tests/view/678c4c8bf4dc0a8b4ae8db5c
LLR: 2.97 (-2.94,2.94) <0.00,2.00>
Total: 273408 W: 71089 L: 70415 D: 131904
Ptnml(0-2): 937, 32466, 69229, 33130, 942
Passed LTC:
https://tests.stockfishchess.org/tests/view/678ccabdf4dc0a8b4ae8dd7a
LLR: 2.95 (-2.94,2.94) <0.50,2.50>
Total: 148614 W: 38138 L: 37584 D: 72892
Ptnml(0-2): 97, 16450, 40689, 16944, 127
closes https://github.com/official-stockfish/Stockfish/pull/5809
Bench: 1582867
2025-01-25 20:27:11 +01:00
Stefan Geschwentner and Joost VandeVondele
59c578ad28
Add move count based reduction.
...
Do less reduction which is linear increasing with move count (factor = 64).
Passed STC:
LLR: 2.94 (-2.94,2.94) <0.00,2.00>
Total: 191488 W: 49982 L: 49432 D: 92074
Ptnml(0-2): 731, 22523, 48614, 23217, 659
https://tests.stockfishchess.org/tests/view/678d0b29d63764e34db4904b
Passed LTC:
LLR: 2.94 (-2.94,2.94) <0.50,2.50>
Total: 90582 W: 23150 L: 22717 D: 44715
Ptnml(0-2): 73, 9936, 24822, 10405, 55
https://tests.stockfishchess.org/tests/view/678d347cd63764e34db4916f
closes https://github.com/official-stockfish/Stockfish/pull/5807
Bench: 1803474
2025-01-25 20:23:33 +01:00
Joost VandeVondele
8e3e22b3d4
Replace depth increase condition with !opponentWorsening
...
Passed simplification STC
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 220544 W: 57417 L: 57399 D: 105728
Ptnml(0-2): 816, 26554, 55540, 26520, 842
https://tests.stockfishchess.org/tests/view/678970e38082388fa0cbfe02
Passed simplification LTC
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 132600 W: 33868 L: 33760 D: 64972
Ptnml(0-2): 126, 14770, 36390, 14898, 116
https://tests.stockfishchess.org/tests/view/678accabc00c743bc9e9fc7f
closes https://github.com/official-stockfish/Stockfish/pull/5798
bench 1632964
2025-01-18 21:33:17 +01:00
Daniel Monroe and Joost VandeVondele
f00d91f8ac
Refine probcut
...
Allow probcut for depth 3 and disable deeper verification search when it would simply repeat the qsearch.
Passed STC
LLR: 2.93 (-2.94,2.94) <0.00,2.00>
Total: 283232 W: 74450 L: 73760 D: 135022
Ptnml(0-2): 1052, 33780, 71349, 34296, 1139
https://tests.stockfishchess.org/tests/view/67843b58460e2910c51ddcc6
Passed LTC
LLR: 2.94 (-2.94,2.94) <0.50,2.50>
Total: 339654 W: 86845 L: 85893 D: 166916
Ptnml(0-2): 298, 37734, 92802, 38704, 289
https://tests.stockfishchess.org/tests/view/678aa722c00c743bc9e9face
closes https://github.com/official-stockfish/Stockfish/pull/5797
bench 1288648
2025-01-18 21:27:42 +01:00
Shawn Xu and Joost VandeVondele
738ac2a100
tuned TM values
...
Tuned 70k games at 240+2.4 th 2: https://tests.stockfishchess.org/tests/view/6783b1b16ddf09c0b4b703f5
Failed STC:
LLR: -2.93 (-2.94,2.94) <0.00,2.00>
Total: 491872 W: 128260 L: 127804 D: 235808
Ptnml(0-2): 1579, 55449, 131572, 55609, 1727
https://tests.stockfishchess.org/tests/view/6785a045460e2910c51de4b8
Passed LTC:
LLR: 2.94 (-2.94,2.94) <0.00,2.00>
Total: 154824 W: 39315 L: 38874 D: 76635
Ptnml(0-2): 110, 15809, 45147, 16222, 124
https://tests.stockfishchess.org/tests/view/678ac722c00c743bc9e9fc35
Passed VLTC:
LLR: 2.95 (-2.94,2.94) <0.50,2.50>
Total: 77404 W: 19825 L: 19452 D: 38127
Ptnml(0-2): 18, 7262, 23765, 7643, 14
https://tests.stockfishchess.org/tests/view/678b2a98c00c743bc9ea048c
closes https://github.com/official-stockfish/Stockfish/pull/5796
No functional change
2025-01-18 21:24:01 +01:00
FauziAkram and Joost VandeVondele
c94bcf62e4
Moving up the if position is or has been on the PV reduction
...
Passed STC:
LLR: 2.93 (-2.94,2.94) <0.00,2.00>
Total: 29664 W: 7880 L: 7570 D: 14214
Ptnml(0-2): 93, 3487, 7390, 3741, 121
https://tests.stockfishchess.org/tests/view/678ac957c00c743bc9e9fc3f
Passed LTC:
LLR: 2.94 (-2.94,2.94) <0.50,2.50>
Total: 81354 W: 20903 L: 20487 D: 39964
Ptnml(0-2): 66, 9003, 22123, 9419, 66
https://tests.stockfishchess.org/tests/view/678ad359c00c743bc9e9fcfa
closes https://github.com/official-stockfish/Stockfish/pull/5794
Bench: 1414638
2025-01-18 21:18:48 +01:00
Michael Chaly and Joost VandeVondele
7701d0b3c4
Introduce one more continuation history
...
This one is counter counter counter history - with really low update value and divided by 3 in movepicker unlike the other ones.
Passed STC:
https://tests.stockfishchess.org/tests/view/67861495460e2910c51de720
LLR: 2.93 (-2.94,2.94) <0.00,2.00>
Total: 44352 W: 11699 L: 11370 D: 21283
Ptnml(0-2): 156, 5098, 11361, 5383, 178
Passed LTC:
https://tests.stockfishchess.org/tests/view/6786e89e3b8f206a2696b646
LLR: 2.94 (-2.94,2.94) <0.50,2.50>
Total: 432660 W: 110355 L: 109207 D: 213098
Ptnml(0-2): 381, 48214, 118039, 49268, 428
closes https://github.com/official-stockfish/Stockfish/pull/5792
Bench: 1491837
2025-01-18 21:04:55 +01:00
Nonlinear2 and Joost VandeVondele
b392ac76db
Increase history bonus of TT moves
...
Passed STC:
https://tests.stockfishchess.org/tests/view/678807653b8f206a2696b78b
LLR: 2.96 (-2.94,2.94) <0.00,2.00>
Total: 42208 W: 11113 L: 10783 D: 20312
Ptnml(0-2): 148, 4919, 10651, 5227, 159
Passed LTC:
https://tests.stockfishchess.org/tests/view/6788a8463b8f206a2696b956
LLR: 2.94 (-2.94,2.94) <0.50,2.50>
Total: 122886 W: 31454 L: 30952 D: 60480
Ptnml(0-2): 105, 13567, 33619, 14025, 127
closes https://github.com/official-stockfish/Stockfish/pull/5791
Bench: 1760081
2025-01-18 21:02:04 +01:00
Shawn Xu and Joost VandeVondele
ccbd060b01
simplify razoring
...
Passed Non-regression STC:
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 93056 W: 24215 L: 24054 D: 44787
Ptnml(0-2): 364, 11085, 23470, 11244, 365
https://tests.stockfishchess.org/tests/view/67883a5d3b8f206a2696b804
Passed Non-regression LTC:
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 66564 W: 16971 L: 16794 D: 32799
Ptnml(0-2): 56, 7403, 18192, 7570, 61
https://tests.stockfishchess.org/tests/view/6789ffa78082388fa0cbfe95
closes https://github.com/official-stockfish/Stockfish/pull/5788
bench 1500649
2025-01-18 20:54:55 +01:00
Daniel Monroe and Joost VandeVondele
a944f08225
retroactive reduction decrease if eval improves
...
If the previous reduction was large but the static eval improves then increase the search depth. This patch looks at the next node when calculating the reduction, something I don't think has been done before and which can probably used for further elo gaining patches.
Passed STC
LLR: 2.93 (-2.94,2.94) <0.00,2.00>
Total: 55936 W: 14813 L: 14462 D: 26661
Ptnml(0-2): 220, 6565, 14094, 6822, 267
https://tests.stockfishchess.org/tests/view/67845b70460e2910c51ddcff
Passed LTC
LLR: 2.94 (-2.94,2.94) <0.50,2.50>
Total: 189468 W: 48411 L: 47773 D: 93284
Ptnml(0-2): 180, 20801, 52131, 21445, 177
https://tests.stockfishchess.org/tests/view/6784e2cb460e2910c51ddf86
closes https://github.com/official-stockfish/Stockfish/pull/5785
bench: 1512884
2025-01-18 20:50:04 +01:00
Disservin and Joost VandeVondele
4c2241089d
Remove addition of 1ms to all timestamps
...
The +1 was a quick fix to avoid the division by zero, a more correct approach is to use 1ms as the minimum reported timestamp to avoid a division by zero.
Later timestamps no longer include an additional 1ms.
closes https://github.com/official-stockfish/Stockfish/pull/5778
No functional change
2025-01-18 20:32:00 +01:00
Daniel Monroe and Joost VandeVondele
5868b4cb58
remove eval== staticeval check in fut pruning
...
Simplify corrplexity in futility margin
Don't check that staticEval == eval when applying the corrplexity-based adjustment in futility pruning.
Passed Simplification STC
LLR: 2.96 (-2.94,2.94) <-1.75,0.25>
Total: 121760 W: 31640 L: 31512 D: 58608
Ptnml(0-2): 349, 14400, 31289, 14458, 384
https://tests.stockfishchess.org/tests/view/6780c4109168c8bf30927777
Passed Simplification LTC
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 134772 W: 34245 L: 34140 D: 66387
Ptnml(0-2): 94, 14869, 37350, 14984, 89
https://tests.stockfishchess.org/tests/view/6782d6ea6ddf09c0b4b6dd36
closes https://github.com/official-stockfish/Stockfish/pull/5776
Bench: 1487627
2025-01-18 20:25:59 +01:00
Carlos Esparza and Joost VandeVondele
aaafaaecf2
prefetch in do_move()
...
this allows removing Position::key_after()
STC
LLR: 2.93 (-2.94,2.94) <-1.75,0.25>
Total: 24960 W: 6556 L: 6336 D: 12068
Ptnml(0-2): 59, 2554, 7056, 2730, 81
LTC
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 115080 W: 29319 L: 29204 D: 56557
Ptnml(0-2): 51, 10736, 35864, 10825, 64
STC with 2MB hash
LLR: 3.04 (-2.94,2.94) <-1.75,0.25>
Total: 182176 W: 46998 L: 46932 D: 88246
Ptnml(0-2): 526, 19711, 50544, 19785, 522
LTC with 8MB hash
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 441180 W: 111557 L: 111746 D: 217877
Ptnml(0-2): 229, 39698, 140929, 39501, 233
closes https://github.com/official-stockfish/Stockfish/pull/5770
bench: 1379150
2025-01-18 20:19:53 +01:00
FauziAkram and Disservin
c085670b84
Increase the depth margin
...
Tested at VVLTC against the passed patches.
Test 1 against PR5764
LLR: 2.94 (-2.94,2.94) <0.50,2.50>
Total: 167260 W: 43053 L: 42521 D: 81686
Ptnml(0-2): 7, 15272, 52542, 15800, 9
https://tests.stockfishchess.org/tests/view/6782ef196ddf09c0b4b6e780
Test 2 against PR5765
LLR: 2.94 (-2.94,2.94) <0.00,2.00>
Total: 300012 W: 77364 L: 76771 D: 145877
Ptnml(0-2): 22, 27555, 94256, 28154, 19
https://tests.stockfishchess.org/tests/view/678366446ddf09c0b4b7028c
closes https://github.com/official-stockfish/Stockfish/pull/5768
Bench: 1379150
2025-01-12 21:36:39 +01:00
Viren6 and Disservin
e2612f9a29
Introduce Correction History Quad Extensions
...
Also modifies the double and triple extension margins with the correction history adjustment.
STC Elo Estimate:
Elo: -4.40 ± 1.4 (95%) LOS: 0.0%
Total: 60000 W: 15230 L: 15990 D: 28780
Ptnml(0-2): 264, 7495, 15168, 6883, 190
nElo: -8.48 ± 2.8 (95%) PairsRatio: 0.91
https://tests.stockfishchess.org/tests/view/6783a3786ddf09c0b4b703a1
Passed 1st VVLTC:
LLR: 2.95 (-2.94,2.94) <0.00,2.00>
Total: 35736 W: 9354 L: 9088 D: 17294
Ptnml(0-2): 4, 3191, 11212, 3457, 4
https://tests.stockfishchess.org/tests/view/6783a3336ddf09c0b4b7039b
Passed 2nd VVLTC:
LLR: 2.95 (-2.94,2.94) <0.50,2.50>
Total: 36394 W: 9515 L: 9225 D: 17654
Ptnml(0-2): 1, 3271, 11364, 3559, 2
https://tests.stockfishchess.org/tests/view/678395e26ddf09c0b4b70345
closes https://github.com/official-stockfish/Stockfish/pull/5767
Bench: 1567166
2025-01-12 21:36:39 +01:00
Muzhen Gaming and Disservin
93edf7a74c
VVLTC Search Tune
...
Values were tuned with 118k VVLTC games.
Tested against #5764 .
Passed VVLTC 1st sprt:
https://tests.stockfishchess.org/tests/view/678331226ddf09c0b4b6fd78
LLR: 2.94 (-2.94,2.94) <0.00,2.00>
Total: 43556 W: 11219 L: 10942 D: 21395
Ptnml(0-2): 2, 3975, 13549, 4248, 4
Passed VVLTC 2nd sprt:
https://tests.stockfishchess.org/tests/view/67834aa06ddf09c0b4b6fe34
LLR: 2.95 (-2.94,2.94) <0.50,2.50>
Total: 37150 W: 9577 L: 9285 D: 18288
Ptnml(0-2): 2, 3344, 11593, 3632, 4
closes https://github.com/official-stockfish/Stockfish/pull/5765
Bench: 1258128
2025-01-12 21:36:31 +01:00
Michael Chaly and Disservin
8b32e4825f
Make IIR less aggressive
...
This patch is an elo gaining simplification which gains elo at longer time controls.
Patch disallows IIR for cutNodes with existing tt moves as well as makes
IIR for pv nodes less aggressive, basiclally confirming suspected
scaling patterns for this heuristic.
Result of 50k games STC run:
https://tests.stockfishchess.org/tests/view/678304676ddf09c0b4b6f9f9
Elo: -2.93 ± 1.6 (95%) LOS: 0.0%
Total: 50000 W: 12718 L: 13140 D: 24142
Ptnml(0-2): 189, 6087, 12835, 5735, 154
nElo: -5.71 ± 3.0 (95%) PairsRatio: 0.94
Passed VVLTC SPRT with STC bounds:
https://tests.stockfishchess.org/tests/view/6782eb1a6ddf09c0b4b6e6b0
LLR: 2.95 (-2.94,2.94) <0.00,2.00>
Total: 150292 W: 38868 L: 38458 D: 72966
Ptnml(0-2): 19, 13890, 46907, 14322, 8
Passed VVLTC SPRT with LTC bounds:
https://tests.stockfishchess.org/tests/view/6782d8d96ddf09c0b4b6df18
LLR: 2.95 (-2.94,2.94) <0.50,2.50>
Total: 153388 W: 39791 L: 39285 D: 74312
Ptnml(0-2): 13, 13924, 48311, 14436, 10
closes https://github.com/official-stockfish/Stockfish/pull/5763
Bench: 1507606
2025-01-12 21:25:36 +01:00
mstembera and Disservin
b84c8807a3
Optimize attackers_to()
...
https://tests.stockfishchess.org/tests/view/6782decb6ddf09c0b4b6e1b0
LLR: 2.93 (-2.94,2.94) <0.00,2.00>
Total: 105920 W: 27571 L: 27181 D: 51168
Ptnml(0-2): 284, 10808, 30403, 11164, 301
- If we only need to know if attackers exist we can skip some
calculations.
- Also calculating slider/magic attackers first is better because the
double lookup is slow due to memory latency.
- I also included a couple of very minor cleanups in search that
probably don't warrant their own PR but I can open separately if
that's better.
closes https://github.com/official-stockfish/Stockfish/pull/5762
No functional change
2025-01-12 21:24:44 +01:00
Shawn Xu and Disservin
921361829a
Simplify away capthist bonus in Probcut
...
The explicit bonus has been obsoleted with the introduction of #5695
Passed Non-regression STC:
LLR: 2.95 (-2.94,2.94) <-1.75,0.25>
Total: 132832 W: 34519 L: 34403 D: 63910
Ptnml(0-2): 430, 15754, 33931, 15872, 429
https://tests.stockfishchess.org/tests/view/678158c49168c8bf30927834
Passed Non-regression LTC:
LLR: 2.95 (-2.94,2.94) <-1.75,0.25>
Total: 123492 W: 31426 L: 31309 D: 60757
Ptnml(0-2): 79, 13705, 34051, 13842, 69
https://tests.stockfishchess.org/tests/view/6782b07e6ddf09c0b4b6dbb7
closes https://github.com/official-stockfish/Stockfish/pull/5761
Bench: 1180439
2025-01-12 21:23:29 +01:00
Shawn Xu and Disservin
28c07fb456
Simplify Probcut Condition
...
Rebased and properly guarded #5720
Passed Non-regression STC:
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 179616 W: 45764 L: 45706 D: 88146
Ptnml(0-2): 125, 19665, 50162, 19739, 117
https://tests.stockfishchess.org/tests/view/677590531a2f267f20548b82
Passed Non-regression LTC:
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 445728 W: 113467 L: 113682 D: 218579
Ptnml(0-2): 331, 49226, 123900, 49141, 266
https://tests.stockfishchess.org/tests/view/67734f351a2f267f205489d9
closes https://github.com/official-stockfish/Stockfish/pull/5758
Bench: 1180421
2025-01-12 21:18:39 +01:00
pb00067 and Disservin
c88a5b3950
Simplify away hint for common parent position at probcut
...
Since it's introduction at probcut step the nnue network has changed
substantially and now it no longer seems useful.
Passed non-regression test at STC
https://tests.stockfishchess.org/tests/view/675fe27986d5ee47d9542d86
LLR: 2.93 (-2.94,2.94) <-1.75,0.25>
Total: 118656 W: 30732 L: 30609 D: 57315
Ptnml(0-2): 319, 12767, 33049, 12858, 335
N.B.: It may be useful to reintroduce it here at probcut if we know that
a node that was cut away previously now has to be explored. Exploring
new ground will deliver no tt-hits so in this case the hint for common
parent position might be useful.
No functional change
2025-01-12 21:15:35 +01:00
Shawn Xu and Disservin
403a5e100b
Simplify Fail-Low Bonus
...
Passed Non-regression STC:
LLR: 2.92 (-2.94,2.94) <-1.75,0.25>
Total: 66592 W: 17426 L: 17239 D: 31927
Ptnml(0-2): 208, 7812, 17109, 7919, 248
https://tests.stockfishchess.org/tests/view/6774e1711a2f267f20548b22
Passed Non-regression LTC:
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 179616 W: 45764 L: 45706 D: 88146
Ptnml(0-2): 125, 19665, 50162, 19739, 117
https://tests.stockfishchess.org/tests/view/677590531a2f267f20548b82
closes https://github.com/official-stockfish/Stockfish/pull/5751
Bench: 1310158
2025-01-12 21:14:11 +01:00
Shawn Xu and Disservin
5370c3035e
Refactor Stats Array
...
* Limit use of `StatsEntry` wrapper to arithmetic types
* Generalize `Stats` to `MultiArray` by discarding the template parameter `D`
* Allow `MultiArray::fill` to take any type assignable to element type
* Remove now-unused operator overloads on `StatsEntry`
closes https://github.com/official-stockfish/Stockfish/pull/5750
No functional change
2025-01-12 21:13:14 +01:00
Shawn Xu and Disservin
d1a1ff4f17
Simplify Razoring
...
Passed Non-regression STC:
LLR: 2.95 (-2.94,2.94) <-1.75,0.25>
Total: 95584 W: 24906 L: 24750 D: 45928
Ptnml(0-2): 285, 11227, 24632, 11343, 305
https://tests.stockfishchess.org/tests/view/675e0ed286d5ee47d95429ee
Passed Non-regression LTC:
LLR: 2.96 (-2.94,2.94) <-1.75,0.25>
Total: 191292 W: 48637 L: 48589 D: 94066
Ptnml(0-2): 97, 21061, 53276, 21121, 91
https://tests.stockfishchess.org/tests/view/675f08c686d5ee47d9542be3
closes https://github.com/official-stockfish/Stockfish/pull/5724
Bench: 1286274
2025-01-12 21:11:27 +01:00
Daniel Monroe and Disservin
c76c179361
Remove non-functional std::min()
...
closes https://github.com/official-stockfish/Stockfish/pull/5749
No functional change
2025-01-06 00:48:05 +01:00
Daniel Monroe and Disservin
1611b9c940
Corrplexity for futility pruning
...
Add corrhist-based term to futility margin
Inspired by a recent patch of Shawn Xu, this tweak increases the margin
over beta needed to futility prune based on the correction history, with
an offset.
Passed STC
LLR: 2.97 (-2.94,2.94) <0.00,2.00>
Total: 545504 W: 141957 L: 140885 D: 262662
Ptnml(0-2): 1829, 64226, 139551, 65336, 1810
https://tests.stockfishchess.org/tests/view/67634a8386d5ee47d95439db
Passed LTC
LLR: 2.94 (-2.94,2.94) <0.50,2.50>
Total: 125994 W: 32199 L: 31695 D: 62100
Ptnml(0-2): 97, 13742, 34798, 14280, 80
https://tests.stockfishchess.org/tests/view/6765cf9986d5ee47d9544217
closes https://github.com/official-stockfish/Stockfish/pull/5748
Bench: 999324
2025-01-06 00:46:30 +01:00
Shawn Xu and Disservin
5488dd2f91
Update Copyright Year
...
closes https://github.com/official-stockfish/Stockfish/pull/5747
No functional change
2025-01-06 00:45:28 +01:00
Shawn Xu and Disservin
78b5733939
Simplify post-lmr conthist bonus
...
Passed Non-regression STC:
LLR: 2.95 (-2.94,2.94) <-1.75,0.25>
Total: 49184 W: 12735 L: 12528 D: 23921
Ptnml(0-2): 134, 5746, 12647, 5909, 156
https://tests.stockfishchess.org/tests/view/6765cd2e86d5ee47d954420e
Passed Non-regression LTC:
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 177270 W: 45227 L: 45166 D: 86877
Ptnml(0-2): 132, 19498, 49302, 19583, 120
https://tests.stockfishchess.org/tests/view/676721fd86d5ee47d9544489
closes https://github.com/official-stockfish/Stockfish/pull/5734
Bench: 1042099
2025-01-06 00:43:37 +01:00
Michael Chaly and Disservin
03e4cde729
Allow Pv nodes at certain conditions to spawn zero window searches deeper than default
...
In current case it's allowed if there is no best move.
Passed STC:
https://tests.stockfishchess.org/tests/view/67640fd586d5ee47d9543d5a
LLR: 2.96 (-2.94,2.94) <0.00,2.00>
Total: 392480 W: 102038 L: 101192 D: 189250
Ptnml(0-2): 1303, 46287, 100253, 47055, 1342
Passed LTC:
https://tests.stockfishchess.org/tests/view/67671a4686d5ee47d9544476
LLR: 2.94 (-2.94,2.94) <0.50,2.50>
Total: 128616 W: 32941 L: 32433 D: 63242
Ptnml(0-2): 84, 13997, 35634, 14513, 80
closes https://github.com/official-stockfish/Stockfish/pull/5733
Bench: 1095871
2025-01-06 00:40:05 +01:00
Michael Chaly and Disservin
79261bec59
Simplify away reductions adjustment for multithreaded search
...
Seem to no longer bring measurable benefit.
Passed STC SMP simplification:
https://tests.stockfishchess.org/tests/view/6753561a86d5ee47d954151f
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 38000 W: 9864 L: 9656 D: 18480
Ptnml(0-2): 53, 4177, 10320, 4409, 41
Passed LTC SMP simplification:
https://tests.stockfishchess.org/tests/view/6753d75f86d5ee47d9541669
LLR: 2.95 (-2.94,2.94) <-1.75,0.25>
Total: 258674 W: 66314 L: 66335 D: 126025
Ptnml(0-2): 77, 26957, 75303, 26910, 90
Passed 16 threads LTC simplification:
https://tests.stockfishchess.org/tests/view/675a066286d5ee47d9542296
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 221804 W: 56950 L: 56936 D: 107918
Ptnml(0-2): 34, 21491, 67839, 21503, 35
closes https://github.com/official-stockfish/Stockfish/pull/5729
Bench: 1294909
2024-12-22 15:28:15 +01:00
Disservin
4bc2a24245
Workaround for clang-format bug
...
closes https://github.com/official-stockfish/Stockfish/pull/5728
No functional change
2024-12-22 15:27:40 +01:00
Shawn Xu and Disservin
e7e78aa09e
Adjust LMR with correction history
...
A positive constant increase in base reduction is applied to counter the
decrease in average reduction from this tweak.
Passed STC:
LLR: 2.98 (-2.94,2.94) <0.00,2.00>
Total: 109216 W: 28415 L: 27989 D: 52812
Ptnml(0-2): 310, 12848, 27911, 13184, 355
https://tests.stockfishchess.org/tests/view/6760bb0e86d5ee47d9542f26
Passed LTC:
LLR: 2.95 (-2.94,2.94) <0.50,2.50>
Total: 66918 W: 17073 L: 16694 D: 33151
Ptnml(0-2): 33, 7175, 18666, 7550, 35
https://tests.stockfishchess.org/tests/view/6761e10f86d5ee47d95431fa
closes https://github.com/official-stockfish/Stockfish/pull/5727
Bench: 1294909
2024-12-22 15:26:39 +01:00
FauziAkram and Disservin
a04b07265f
Make reductionScale smoother
...
Making the second part of the formula smoother, changing it to a linear
function, increasing steadily as reductionScale increases and at the
same time, it should be a little bit simpler, therefore the
simplification bounds.
Passed STC:
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 83040 W: 21493 L: 21322 D: 40225
Ptnml(0-2): 252, 9848, 21209, 9899, 312
https://tests.stockfishchess.org/tests/view/6762145486d5ee47d9543242
Passed LTC:
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 230124 W: 58485 L: 58478 D: 113161
Ptnml(0-2): 175, 25620, 63484, 25589, 194
https://tests.stockfishchess.org/tests/view/6762d4ef86d5ee47d9543367
closes https://github.com/official-stockfish/Stockfish/pull/5725
Bench: 1204658
2024-12-22 15:24:09 +01:00
Disservin
2dc47e4345
Cleanup Evaluate Calls
...
Makes code a bit easier to read as well.
closes https://github.com/official-stockfish/Stockfish/pull/5722
No functional change
2024-12-22 15:19:53 +01:00
Daniel Monroe and Disservin
ba145332c9
Remove time reduction for recaptures
...
Passed simplification STC
LLR: 2.96 (-2.94,2.94) <-1.75,0.25>
Total: 54016 W: 14098 L: 13902 D: 26016
Ptnml(0-2): 165, 5797, 14919, 5931, 196
https://tests.stockfishchess.org/tests/view/6758a90486d5ee47d954201e
Passed simplification LTC
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 296940 W: 75631 L: 75689 D: 145620
Ptnml(0-2): 145, 28928, 90384, 28866, 147
https://tests.stockfishchess.org/tests/view/6758df7a86d5ee47d9542091
closes https://github.com/official-stockfish/Stockfish/pull/5719
Bench: 1148169
2024-12-22 15:17:55 +01:00
Shawn Xu and Disservin
e770b55f7f
Remove Extraneous Parenthesis
...
No longer needed after
https://github.com/official-stockfish/Stockfish/pull/5667 .
closes https://github.com/official-stockfish/Stockfish/pull/5717
No functional change
2024-12-22 15:17:33 +01:00
Shawn Xu and Disservin
b822fdf2f2
Tune histories
...
Passed STC:
LLR: 2.93 (-2.94,2.94) <0.00,2.00>
Total: 268736 W: 70080 L: 69421 D: 129235
Ptnml(0-2): 831, 31795, 68460, 32448, 834
https://tests.stockfishchess.org/tests/view/6750778886d5ee47d9540e7c
Passed LTC:
LLR: 2.95 (-2.94,2.94) <0.50,2.50>
Total: 505356 W: 129145 L: 127868 D: 248343
Ptnml(0-2): 307, 54901, 140959, 56230, 281
https://tests.stockfishchess.org/tests/view/675367de86d5ee47d9541536
closes https://github.com/official-stockfish/Stockfish/pull/5712
Bench: 1148169
2024-12-22 15:14:22 +01:00
Nonlinear2 and GitHub
cf10644d6e
Fix duplicate code ( #5711 )
...
closes https://github.com/official-stockfish/Stockfish/pull/5711
No functional change
2024-12-08 22:24:29 +01:00
Michael Chaly and Disservin
afaf3a0f2a
Refine statscore for captures
...
Continuation of previous attempts there.
Now instead of using capture history with a static offset also add the
value of the captured piece in the same way at it is used in movepicker.
Passed STC:
https://tests.stockfishchess.org/tests/view/674aa3d386d5ee47d95404aa
LLR: 2.93 (-2.94,2.94) <0.00,2.00>
Total: 116480 W: 30433 L: 29999 D: 56048
Ptnml(0-2): 361, 13720, 29662, 14118, 379
Passed LTC:
https://tests.stockfishchess.org/tests/view/674c4b2d86d5ee47d954073f
LLR: 2.94 (-2.94,2.94) <0.50,2.50>
Total: 133542 W: 34365 L: 33847 D: 65330
Ptnml(0-2): 78, 14585, 36934, 15089, 85
closes https://github.com/official-stockfish/Stockfish/pull/5706
Bench: 934447
2024-12-08 19:57:59 +01:00
Shawn Xu and Disservin
6a8478c6ad
Simplify Prior Capture Countermove Bonus
...
Passed Non-regression STC:
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 184032 W: 47626 L: 47568 D: 88838
Ptnml(0-2): 590, 21808, 47238, 21714, 666
https://tests.stockfishchess.org/tests/view/67412c7686d5ee47d953f743
Passed Non-regression LTC:
LLR: 2.96 (-2.94,2.94) <-1.75,0.25>
Total: 169218 W: 43395 L: 43323 D: 82500
Ptnml(0-2): 302, 18567, 46791, 18655, 294
https://tests.stockfishchess.org/tests/view/6743b7e086d5ee47d953f9a6
closes https://github.com/official-stockfish/Stockfish/pull/5701
Bench: 1130692
2024-12-08 19:55:10 +01:00
pb00067 and Disservin
713000c517
Same weight for black and white nonPawnCorrection history
...
Since we don't have color dependent parameters in NNUE eval, it also has
no sense IMO to have color dependent parameters in correction histories.
Ideally a fixed depth search on a single thread should be determistic,
so delivering the same result (move) if we just flip colors on the
board.
Patch replaces 2 parameters (122 and 185) with just one value 154 (= the avg of the two).
Passed STC-non regression
https://tests.stockfishchess.org/tests/view/6740a63286d5ee47d953f656
LLR: 2.95 (-2.94,2.94) <-1.75,0.25>
Total: 122336 W: 31499 L: 31372 D: 59465
Ptnml(0-2): 336, 14535, 31301, 14658, 338
Passed LTC-non regression
https://tests.stockfishchess.org/tests/view/67419bae86d5ee47d953f7b6
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 101400 W: 25870 L: 25731 D: 49799
Ptnml(0-2): 78, 11109, 28166, 11290, 57
closes https://github.com/official-stockfish/Stockfish/pull/5698
Bench: 1215483
2024-12-08 19:52:16 +01:00
FauziAkram and Disservin
d5a36a3c92
Simplify probCutBeta formula
...
After recent changes to the improving definition, seems like there is no
need anymore to keep opponentWorsening in the probCutBeta formula.
Passed STC:
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 214272 W: 55566 L: 55541 D: 103165
Ptnml(0-2): 620, 25540, 54817, 25513, 646
https://tests.stockfishchess.org/tests/view/6735243d86d5ee47d953eaea
Passed LTC:
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 126708 W: 32329 L: 32216 D: 62163
Ptnml(0-2): 68, 13986, 35123, 14119, 58
https://tests.stockfishchess.org/tests/view/67393cf686d5ee47d953ef99
closes https://github.com/official-stockfish/Stockfish/pull/5697
Bench: 983067
2024-12-08 19:49:01 +01:00
Nonlinear2 and Disservin
da82942b54
Add functions to check for decisive scores
...
Thanks to peregrineshahin and robbyrobbyrob for their suggestions.
closes https://github.com/official-stockfish/Stockfish/pull/5696
No functional change
2024-12-08 19:48:49 +01:00
Nonlinear2 and Disservin
57e06be71f
Add functions to check for decisive scores
...
Thanks to peregrineshahin and robbyrobbyrob for their suggestions.
closes https://github.com/official-stockfish/Stockfish/pull/5696
No functional change
2024-12-08 19:47:02 +01:00
Daniel Monroe and Disservin
70bb317afe
Bonus for a prior capture that causes a fail low.
...
This tweak adds a bonus equal to twice the stat_bonus for the current depth for a prior capture that caused a fail high, similar to the prior countermove bonus we currently have.
Passed STC
https://tests.stockfishchess.org/tests/view/673bc14b86d5ee47d953f1f2
LLR: 2.95 (-2.94,2.94) <0.00,2.00>
Total: 105824 W: 27538 L: 27118 D: 51168
Ptnml(0-2): 358, 12370, 27024, 12814, 346
Passed LTC
https://tests.stockfishchess.org/tests/view/673ccbff86d5ee47d953f2d9
LLR: 2.94 (-2.94,2.94) <0.50,2.50>
Total: 134502 W: 34340 L: 33820 D: 66342
Ptnml(0-2): 102, 14634, 37229, 15214, 72
closes https://github.com/official-stockfish/Stockfish/pull/5695
Bench: 1107054
2024-12-08 19:46:07 +01:00
Daniel Monroe and Disservin
fb6be17ad4
Simplify statscore at captures
...
Simplify statscores for captures, setting them to 0
A recent tweak of Vizvezdenec finds substantial elo gain from giving
captures a separate statscore, which is used mainly for reductions. The
idea is that the old combination of quiet histories was inappropriate
and that a value based on the capture history is more suitable. This
simplification sets the statscore for captures to 0, suggesting that the
elo gain came from rectifying the quiet history/capture mismatch.
Passed STC (against a slightly older version of Viz's patch)
https://tests.stockfishchess.org/tests/view/673ac6e286d5ee47d953f0ec
LR: 2.96 (-2.94,2.94) <-1.75,0.25>
Total: 57312 W: 14872 L: 14672 D: 27768
Ptnml(0-2): 152, 6761, 14649, 6923, 171
Passed LTC (against Viz's newest patch)
https://tests.stockfishchess.org/tests/view/673cd00686d5ee47d953f2db
LLR: 2.95 (-2.94,2.94) <-1.75,0.25>
Total: 88236 W: 22510 L: 22358 D: 43368
Ptnml(0-2): 70, 9530, 24745, 9724, 49
closes https://github.com/official-stockfish/Stockfish/pull/5691
Bench: 959947
2024-11-22 23:34:15 +01:00
Shawn Xu and Disservin
4fcd78ceb4
Simplify Probcut Bonus
...
Passed STC:
LLR: 2.99 (-2.94,2.94) <-1.75,0.25>
Total: 172288 W: 44656 L: 44580 D: 83052
Ptnml(0-2): 507, 20650, 43782, 20670, 535
https://tests.stockfishchess.org/tests/view/673b74f986d5ee47d953f1a3
Passed LTC:
LLR: 2.94 (-2.94,2.94) <-1.75,0.25>
Total: 94596 W: 24098 L: 23953 D: 46545
Ptnml(0-2): 57, 10322, 26393, 10471, 55
https://tests.stockfishchess.org/tests/view/673d191886d5ee47d953f337
closes https://github.com/official-stockfish/Stockfish/pull/5688
Bench: 1031022
2024-11-22 23:31:34 +01:00
Michael Chaly and Disservin
cd3c13a883
Further tweak statscore for captures
...
Even lower offset.
Passed STC:
https://tests.stockfishchess.org/tests/view/673a66d786d5ee47d953f070
LLR: 2.94 (-2.94,2.94) <0.00,2.00>
Total: 63776 W: 16570 L: 16216 D: 30990
Ptnml(0-2): 178, 7371, 16478, 7641, 220
Passed LTC:
https://tests.stockfishchess.org/tests/view/673b2e2a86d5ee47d953f14b
LLR: 2.95 (-2.94,2.94) <0.50,2.50>
Total: 156960 W: 39999 L: 39435 D: 77526
Ptnml(0-2): 96, 16965, 43803, 17511, 105
closes https://github.com/official-stockfish/Stockfish/pull/5686
Bench: 867931
2024-11-22 23:29:37 +01:00
Guenther Demetz and Disservin
d29c8bd5d4
Rewrite of 'Adjust correction history' condition
...
Current condition is convoluted and hard to understand because of
several negations. Also added 2 comments to make the concept behind the
condition better understandable.
closes https://github.com/official-stockfish/Stockfish/pull/5685
No functional change
2024-11-22 23:24:15 +01:00
FauziAkram and Disservin
0282edc0b0
Simplify bonus formula
...
Give full bonus instead of half.
Passed STC:
LLR: 2.95 (-2.94,2.94) <-1.75,0.25>
Total: 23872 W: 6254 L: 6018 D: 11600
Ptnml(0-2): 80, 2691, 6152, 2939, 74
https://tests.stockfishchess.org/tests/view/673b709686d5ee47d953f19d
Passed LTC:
LLR: 2.95 (-2.94,2.94) <-1.75,0.25>
Total: 42894 W: 10924 L: 10725 D: 21245
Ptnml(0-2): 30, 4592, 12011, 4777, 37
https://tests.stockfishchess.org/tests/view/673bb50386d5ee47d953f1eb
closes https://github.com/official-stockfish/Stockfish/pull/5683
Bench: 836558
2024-11-22 23:23:42 +01:00
