14f83ad7b9
Move public search/qsearch interface from namespace Learner to namespace Search
2020-10-15 20:37:03 +09:00
0494adeb2c
Move nnue evaluation stuff from evaluate.h to nnue/evaluate_nnue.h
2020-10-15 20:37:03 +09:00
a8b502a975
Merge remote-tracking branch 'remotes/origin/master'
...
Bench: 3618595
2020-09-29 17:09:14 +08:00
6f0aa186d8
Tweak reduction formula.
...
Replace log(i) with log(i + 0.25 * log(i)). This increases especially for low values the reductions. But for bigger values there are nearly no changes.
STC:
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 49640 W: 5505 L: 5289 D: 38846
Ptnml(0-2): 270, 4074, 15924, 4274, 278
https://tests.stockfishchess.org/tests/view/5f71f04d3b22d6afa5069478
LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 43856 W: 2209 L: 2021 D: 39626
Ptnml(0-2): 32, 1776, 18128, 1956, 36
https://tests.stockfishchess.org/tests/view/5f7232ee3b22d6afa50699a2
closes https://github.com/official-stockfish/Stockfish/pull/3163
Bench: 3555769
2020-09-28 22:34:25 +02:00
a5e68d9b25
Adjust null move pruning constants
...
Idea is that division by fraction of 2 is slightly faster than by other numbers so parameters are adjusted in a way that division in null move pruning depth reduction features dividing by 256 instead of dividing by 213.
Other than this patch is almost non-functional - difference starts to exist by depth 133.
passed STC
https://tests.stockfishchess.org/tests/view/5f70dd943b22d6afa50693c5
LLR: 2.95 (-2.94,2.94) {-0.25,1.25}
Total: 57048 W: 6616 L: 6392 D: 44040
Ptnml(0-2): 304, 4583, 18531, 4797, 309
passed LTC
https://tests.stockfishchess.org/tests/view/5f7180db3b22d6afa506941f
LLR: 2.95 (-2.94,2.94) {0.25,1.25}
Total: 45960 W: 2419 L: 2229 D: 41312
Ptnml(0-2): 43, 1779, 19137, 1987, 34
closes https://github.com/official-stockfish/Stockfish/pull/3159
bench 3789924
2020-09-28 22:22:54 +02:00
5be8b573be
Merge remote-tracking branch 'remotes/origin/master' into trainer
2020-09-23 19:02:27 +08:00
411adab149
Merge remote-tracking branch 'remotes/nodchip/master' into trainer
2020-09-23 18:29:30 +08:00
3d5b2c8a51
Increase reductions with the number of threads
...
Passed STC with 8 threads:
LLR: 2.92 (-2.94,2.94) {-0.25,1.25}
Total: 13520 W: 1135 L: 1012 D: 11373
Ptnml(0-2): 39, 815, 4929, 938, 39
https://tests.stockfishchess.org/tests/view/5f68e274ded68c240be73f41
Passed LTC with 8 threads:
LLR: 2.96 (-2.94,2.94) {0.25,1.25}
Total: 48384 W: 2183 L: 1994 D: 44207
Ptnml(0-2): 28, 1777, 20402, 1948, 37
https://tests.stockfishchess.org/tests/view/5f68f068ded68c240be747e9
closes https://github.com/official-stockfish/Stockfish/pull/3142
No functional change (for one thread)
2020-09-22 22:45:02 +02:00
9a64e737cf
Small cleanups 12
...
- Clean signature of functions in namespace NNUE
- Add comment for countermove based pruning
- Remove bestMoveCount variable
- Add const qualifier to kpp_board_index array
- Fix spaces in get_best_thread()
- Fix indention in capture LMR code in search.cpp
- Rename TtmemDeleter to LargePageDeleter
Closes https://github.com/official-stockfish/Stockfish/pull/3063
No functional change
2020-09-21 10:41:10 +02:00
8559c43914
Simplify reduced depth search
...
Simplification in reduced depth search.
STC https://tests.stockfishchess.org/tests/view/5f64c72fbb0cae038ca8f531
LLR: 2.94 (-2.94,2.94) {-1.25,0.25}
Total: 28320 W: 3475 L: 3359 D: 21486
Ptnml(0-2): 170, 2485, 8773, 2523, 209
LTC https://tests.stockfishchess.org/tests/view/5f650cfabb0cae038ca8f585
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 58392 W: 3354 L: 3285 D: 51753
Ptnml(0-2): 74, 2826, 23336, 2877, 83
closes https://github.com/official-stockfish/Stockfish/pull/3139
bench: 4201295
2020-09-21 07:47:41 +02:00
e8472b5fbe
Fix races in gensfen as detected with thread sanitizer.
...
RootInTB was an incorrectly shared global, probably leading to wrong scoreing
Minor:
setting TB global state from input by all threads (all threads write same values)
setting Limits global state by all threads (idem)
thread counting for finalization
CI can be enabled once races are fixed in the learner, manually goes like:
```
make clean && make -j2 ARCH=x86-64-modern sanitize=thread optimize=no debug=yes build
../tests/instrumented_learn.sh --sanitizer-thread
```
Needs some review.
2020-09-19 10:08:44 +09:00
26f63fe741
Merge remote-tracking branch 'remotes/origin/master' into trainer
2020-09-19 03:38:37 +08:00
a47a3bfc7c
Merge remote-tracking branch 'remotes/nodchip/master' into trainer
2020-09-19 02:14:17 +08:00
64a63464d7
Simplify futility pruning for captures
...
STC https://tests.stockfishchess.org/tests/view/5f61f0e4b91f2ec371e429c2
LLR: 2.94 (-2.94,2.94) {-1.25,0.25}
Total: 75512 W: 8747 L: 8704 D: 58061
Ptnml(0-2): 440, 6589, 23683, 6576, 468
LTC https://tests.stockfishchess.org/tests/view/5f6215d3912c15f19854b801
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 92912 W: 5030 L: 4992 D: 82890
Ptnml(0-2): 88, 4363, 37532, 4369, 104
closes https://github.com/official-stockfish/Stockfish/pull/3129
bench: 3856086
2020-09-17 07:06:21 +02:00
0ca93c5b94
Remove castling extension
...
STC https://tests.stockfishchess.org/tests/view/5f5fa5348fbc1c8a3f476eca
LLR: 2.94 (-2.94,2.94) {-1.25,0.25}
Total: 38520 W: 4713 L: 4610 D: 29197
Ptnml(0-2): 233, 3486, 11734, 3559, 248
LTC https://tests.stockfishchess.org/tests/view/5f62166a912c15f19854b806
LLR: 2.93 (-2.94,2.94) {-0.75,0.25}
Total: 48024 W: 2673 L: 2600 D: 42751
Ptnml(0-2): 64, 2247, 19316, 2322, 63
closes https://github.com/official-stockfish/Stockfish/pull/3128
bench: 3818400
2020-09-17 07:06:21 +02:00
5f426d8667
Use 2 * bestMoveChanges.
...
NNUE appears to provide a more stable eval than the classic eval,
so the time use dependencies on bestMoveChanges, fallingEval,
etc may need to change to make the best use of available time.
This change doubles the effect of totBestMoveChanges when giving
more time because the choice of best move is unstable.
STC:
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 101928 W: 11995 L: 11698 D: 78235 Elo +0.78
Ptnml(0-2): 592, 8707, 32103, 8936, 626
https://tests.stockfishchess.org/tests/view/5f538a462d02727c56b36cec
LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 186392 W: 10383 L: 9877 D: 166132 Elo +0.81
Ptnml(0-2): 207, 8370, 75539, 8870, 210
https://tests.stockfishchess.org/tests/view/5f54a9712d02727c56b36d5a
closes https://github.com/official-stockfish/Stockfish/pull/3119
Bench 4222126
2020-09-16 20:56:40 +02:00
35ab8254b7
Simplify StatSCore Initialization
...
No need to initialize StatScore at rootNode. Current Logic is redundant because at subsequent levels the grandchildren statScore is initialized to zero.
closes https://github.com/official-stockfish/Stockfish/pull/3122
Non functional change.
2020-09-15 07:19:02 +02:00
30a1bc4c64
Change default value of "PruneAtShallowDepthOnPvNode" so that the bench matches master.
2020-09-13 14:19:30 +02:00
d33e7a9b07
Remove conditional compilation on EVAL_LEARN
2020-09-12 16:19:24 +02:00
94f3cae760
Changed a sentence.
2020-09-10 08:23:21 +09:00
e63b6088ba
Changed a option name more descriptive, "Training" -> "PruneAtShallowDepthOnPvNode". The default value was changed but the default behavior is not changed.
...
Changed to set a global option prune_at_shallow_depth_on_pv_node on a callback function.
2020-09-10 08:19:54 +09:00
675d336ebb
Merge branch 'master' into trainer
2020-09-09 16:08:49 +08:00
82dc68ba9f
Removed #if for USE_GLOBAL_OPTIONS.
2020-09-09 10:26:42 +09:00
d25657c439
Merge branch 'master' into trainer
2020-09-09 08:43:12 +08:00
d21424c8d3
test
2020-09-09 07:31:22 +08:00
3bf418e63f
Fix some uninitialized variables with gensfen
...
fixes valgrind errors as seen with:
```
setoption name Use NNUE value true
isready
gensfen depth 6 loop 10 use_draw_in_training_data_generation 1 eval_limit 32000 output_file_name training_data/training_data.bin use_raw_nnue_eval 0
quit
```
the latter script now runs without valgrind errors on linux
2020-09-07 09:01:17 +09:00
9a063fc3cb
Adjust penalty on refuted early quiet moves
...
This patch changes how previous early moves are penalized in case
search finds a best move. Here, the first quiet move that was not
a transposition table move is penalized.
passed STC
https://tests.stockfishchess.org/tests/view/5f51d839ba100690c5cc5f69
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 10088 W: 1150 L: 997 D: 7941
Ptnml(0-2): 41, 772, 3278, 899, 54
passed LTC
https://tests.stockfishchess.org/tests/view/5f51e435ba100690c5cc5f76
LLR: 2.93 (-2.94,2.94) {0.25,1.25}
Total: 30808 W: 1564 L: 1405 D: 27839
Ptnml(0-2): 19, 1245, 12717, 1404, 19
closes https://github.com/official-stockfish/Stockfish/pull/3106
bench 3983758
2020-09-04 18:52:46 +02:00
2a69611509
LMR Simplification
...
remove reduction at non-check cut nodes for second move at low depths
STC:
LLR: 2.95 (-2.94,2.94) {-1.25,0.25}
Total: 61712 W: 6594 L: 6543 D: 48575
Ptnml(0-2): 293, 5085, 20082, 5070, 326
https://tests.stockfishchess.org/tests/view/5f5007d6ba100690c5cc5ea9
LTC:
LLR: 2.94 (-2.94,2.94) {-0.75,0.25}
Total: 57544 W: 2983 L: 2925 D: 51636
Ptnml(0-2): 47, 2568, 23495, 2604, 58
https://tests.stockfishchess.org/tests/view/5f50c597ba100690c5cc5ef7
closes https://github.com/official-stockfish/Stockfish/pull/3103
Bench: 3952302
2020-09-04 08:00:45 +02:00
d6530f7d49
Simplify singularQuietLMR
...
remove formerPV dependence
STC https://tests.stockfishchess.org/tests/view/5f4cb922ba100690c5cc5d35
LLR: 2.96 (-2.94,2.94) {-1.25,0.25}
Total: 113672 W: 12347 L: 12368 D: 88957
Ptnml(0-2): 566, 9537, 36699, 9420, 614
LTC https://tests.stockfishchess.org/tests/view/5f4e8474ba100690c5cc5e12
LLR: 2.93 (-2.94,2.94) {-0.75,0.25}
Total: 43032 W: 2298 L: 2227 D: 38507
Ptnml(0-2): 45, 1940, 17475, 2011, 45
closes https://github.com/official-stockfish/Stockfish/pull/3102
bench: 3290084
2020-09-04 07:58:13 +02:00
0e1f734b05
Less pruning in qsearch
...
do not prune moves that give discovery checks, even if with negative SSE.
STC https://tests.stockfishchess.org/tests/view/5f4cb5e8ba100690c5cc5d25
LLR: 2.96 (-2.94,2.94) {-0.25,1.25}
Total: 91328 W: 9940 L: 9667 D: 71721
Ptnml(0-2): 491, 7345, 29693, 7670, 465
LTC https://tests.stockfishchess.org/tests/view/5f4dbc2eba100690c5cc5dac
LLR: 2.97 (-2.94,2.94) {0.25,1.25}
Total: 52448 W: 2799 L: 2586 D: 47063
Ptnml(0-2): 53, 2220, 21459, 2445, 47
closes https://github.com/official-stockfish/Stockfish/pull/3098
bench: 4031192
2020-09-04 07:55:41 +02:00
a8bbaa1795
LMR Root Node Simplification
...
Simplify LMR at Root node
STC:
LLR: 2.94 (-2.94,2.94) {-1.25,0.25}
Total: 71520 W: 7649 L: 7614 D: 56257
Ptnml(0-2): 346, 5845, 23349, 5868, 352
https://tests.stockfishchess.org/tests/view/5f4be8c0ba100690c5cc5cbb
LTC:
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 74832 W: 3997 L: 3948 D: 66887
Ptnml(0-2): 77, 3422, 30362, 3485, 70
https://tests.stockfishchess.org/tests/view/5f4c603eba100690c5cc5d0e
closes https://github.com/official-stockfish/Stockfish/pull/3091
Bench: 3624569
2020-09-01 09:31:40 +02:00
61381372ec
Always print an info line before a bestmove
...
if very few nodes are being searched before a bestmove is reported,
an info line might be missing.
fixes https://github.com/official-stockfish/Stockfish/issues/2757
closes https://github.com/official-stockfish/Stockfish/pull/3088
no functional change
2020-09-01 09:26:56 +02:00
a0afe32d16
Use stable sort to make sure bench with TB yields same results everywhere.
...
std::sort() is not stable so different implementations can produce different results:
use the stable version instead.
Observed for '8/6k1/5r2/8/8/8/1K6/Q7 w - - 0 1' yielding different bench results for gcc and MSVC
and 3-4-5 syzygy TB prior to this patch.
closes https://github.com/official-stockfish/Stockfish/pull/3083
No functional change.
2020-09-01 09:25:44 +02:00
e0bafa1911
Update parametes in classical evaluation.
...
Passed STC (NNUE=False):
https://tests.stockfishchess.org/tests/view/5f42edfe5089a564a10d84a0
LLR: 2.96 (-2.94,2.94) {-0.25,1.25}
Total: 13840 W: 2591 L: 2336 D: 8913
Ptnml(0-2): 194, 1453, 3387, 1676, 210
Passed LTC (NNUE=False):
https://tests.stockfishchess.org/tests/view/5f4369795089a564a10d84d8
LLR: 2.95 (-2.94,2.94) {0.25,1.25}
Total: 159744 W: 19430 L: 18850 D: 121464
Ptnml(0-2): 960, 14185, 49030, 14709, 988
closes https://github.com/official-stockfish/Stockfish/pull/3080
bench: 3736029
2020-08-30 14:22:07 +02:00
9b5b9ec9a6
QS Pruning Simplification
...
Remove depth dependence in QS pruning
STC:
LLR: 2.95 (-2.94,2.94) {-1.25,0.25}
Total: 40536 W: 4442 L: 4358 D: 31736
Ptnml(0-2): 209, 3330, 13118, 3390, 221
https://tests.stockfishchess.org/tests/view/5f49035b3def6407861152f9
LTC:
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 97104 W: 5164 L: 5130 D: 86810
Ptnml(0-2): 103, 4478, 39377, 4470, 124
https://tests.stockfishchess.org/tests/view/5f4939d53def640786115322
closes https://github.com/official-stockfish/Stockfish/pull/3077
Bench: 3865238
2020-08-30 14:17:23 +02:00
c02b3a4c7a
Add / remove leaves from search tree ttPv
...
add if previous leaf is in search tree and we didn't find a counter move
else remove the position if the leaf is the last one in search tree.
STC : https://tests.stockfishchess.org/tests/view/5f49203c3def640786115314
LLR: 2.95 (-2.94,2.94) {-0.25,1.25}
Total: 29968 W: 3381 L: 3195 D: 23392
Ptnml(0-2): 146, 2432, 9671, 2560, 175
LTC : https://tests.stockfishchess.org/tests/view/5f494bea3def640786115336
LLR: 2.96 (-2.94,2.94) {0.25,1.25}
Total: 84952 W: 4619 L: 4333 D: 76000
Ptnml(0-2): 86, 3765, 34481, 4065, 79
closes https://github.com/official-stockfish/Stockfish/pull/3075
Bench 3527337
2020-08-30 14:13:16 +02:00
d90d893b5e
Reintroduce depth reduction
...
Reintroduce depth reduction if the position is not in TT.
STC https://tests.stockfishchess.org/tests/view/5f4652e85089a564a10d868c
LLR: 2.97 (-2.94,2.94) {-0.25,1.25}
Total: 40240 W: 4535 L: 4331 D: 31374
Ptnml(0-2): 215, 3276, 12969, 3410, 250
LTC https://tests.stockfishchess.org/tests/view/5f46ca5e5089a564a10d86f3
LLR: 2.93 (-2.94,2.94) {0.25,1.25}
Total: 63096 W: 3426 L: 3188 D: 56482
Ptnml(0-2): 51, 2798, 25645, 2970, 84
closes https://github.com/official-stockfish/Stockfish/pull/3072
bench: 3611906
2020-08-30 14:04:29 +02:00
906c18eb46
Merge branch 'master' of github.com:official-stockfish/Stockfish into nnue-player-merge-2020-08-28
...
# Conflicts:
# README.md
# src/Makefile
# src/search.cpp
# src/types.h
# src/uci.cpp
# src/ucioption.cpp
2020-08-28 11:26:11 +09:00
242a7d9fea
Simplify MCP in QS
...
Simplify moveCount pruning in QS by removing depth dependency.
STC
LLR: 2.94 (-2.94,2.94) {-1.25,0.25}
Total: 42960 W: 4741 L: 4661 D: 33558
Ptnml(0-2): 218, 3574, 13804, 3678, 206
https://tests.stockfishchess.org/tests/view/5f42e3f75089a564a10d8493
LTC
LLR: 2.94 (-2.94,2.94) {-0.75,0.25}
Total: 66672 W: 3563 L: 3508 D: 59601
Ptnml(0-2): 71, 3064, 26996, 3149, 56
https://tests.stockfishchess.org/tests/view/5f4353285089a564a10d84d0
closes https://github.com/official-stockfish/Stockfish/pull/3067
Bench: 4074430
2020-08-26 07:19:35 +02:00
95b8f3f800
Remove Reduce Depth
...
Remove Reduce Depth at PV nodes.
STC:
LLR: 2.94 (-2.94,2.94) {-1.25,0.25}
Total: 56760 W: 6299 L: 6236 D: 44225
Ptnml(0-2): 286, 4843, 18076, 4872, 303
https://tests.stockfishchess.org/tests/view/5f41356087a5c3c63d8f53c9
LTC:
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 17496 W: 954 L: 865 D: 15677
Ptnml(0-2): 13, 768, 7098, 855, 14
https://tests.stockfishchess.org/tests/view/5f41bb7687a5c3c63d8f53f9
closes https://github.com/official-stockfish/Stockfish/pull/3055
Bench: 3555051
2020-08-26 07:16:50 +02:00
843a961a8c
Introduce countermove based pruning for qsearch
...
This patch continues work of previous patch in introducing pruning heuristics in qsearch by analogy to main search, now with countermove based pruning.
Idea is that if move is late enough and is quite check (we do generate them in qsearch) and has bad enough countermove history - prune it.
passed STC
https://tests.stockfishchess.org/tests/view/5f41220287a5c3c63d8f53c5
LLR: 2.93 (-2.94,2.94) {-0.25,1.25}
Total: 35944 W: 4127 L: 3929 D: 27888
Ptnml(0-2): 196, 2970, 11459, 3134, 213
passed LTC
https://tests.stockfishchess.org/tests/view/5f41862f87a5c3c63d8f53e8
LLR: 2.95 (-2.94,2.94) {0.25,1.25}
Total: 138448 W: 7655 L: 7252 D: 123541
Ptnml(0-2): 145, 6247, 56043, 6638, 151
closes https://github.com/official-stockfish/Stockfish/pull/3058
Bench: 3610676
2020-08-24 12:09:04 +02:00
d5f86b6359
Introduce movecount pruning for qsearch()
...
If in quiescence search, we assume that me can prune late moves when:
a) the move ordering count of the move is : moveCount > abs(depth) + 2
b) we are not in check
c) the late move does not give check
d) the late move is not an advanced pawn push
Modification of an original idea by @VoyagerOne.
STC
https://tests.stockfishchess.org/tests/view/5f40581787a5c3c63d8f535f
LLR: 2.96 (-2.94,2.94) {-0.25,1.25}
Total: 132848 W: 14999 L: 14661 D: 103188
Ptnml(0-2): 684, 11242, 42309, 11430, 759
LTC
https://tests.stockfishchess.org/tests/view/5f4226da87a5c3c63d8f5412
LLR: 2.98 (-2.94,2.94) {0.25,1.25}
Total: 12008 W: 678 L: 551 D: 10779
Ptnml(0-2): 8, 485, 4899, 596, 16
closes https://github.com/official-stockfish/Stockfish/pull/3053
Bench: 3749974
2020-08-23 23:27:03 +02:00
5f1843c9cb
Small trivial cleanups
...
closes https://github.com/official-stockfish/Stockfish/pull/2801
No functional change
2020-08-23 01:53:41 +02:00
e64b957274
Simplify away internal iterative deepening
...
Remove the iterative deepening step.
Instead, employ a depth reduction if the position is not in TT and on the PV.
STC https://tests.stockfishchess.org/tests/view/5f3ce6eaa95672ddd56c637e
LLR: 2.97 (-2.94,2.94) {-0.50,1.50}
Total: 41096 W: 4421 L: 4257 D: 32418
Ptnml(0-2): 207, 3259, 13460, 3407, 215
LTC (old) https://tests.stockfishchess.org/tests/view/5f3d7d4fa95672ddd56c640b
LLR: 2.92 (-2.94,2.94) {-1.50,0.50}
Total: 26032 W: 1320 L: 1309 D: 23403
Ptnml(0-2): 22, 1152, 10654, 1169, 19
LTC (new) https://tests.stockfishchess.org/tests/view/5f3e31e0a95672ddd56c6464
LLR: 2.95 (-2.94,2.94) {-0.75,0.25}
Total: 34160 W: 1844 L: 1766 D: 30550
Ptnml(0-2): 33, 1533, 13876, 1599, 39
bench: 3849173
2020-08-21 18:04:14 +02:00
2deb08a529
Reintroduce last captures extension
...
STC:
LLR: 2.93 (-2.94,2.94) {-0.50,1.50}
Total: 34840 W: 3834 L: 3682 D: 27324
Ptnml(0-2): 153, 2767, 11455, 2865, 180
https://tests.stockfishchess.org/tests/view/5f3bb380b38d442594aabefc
LTC:
LLR: 2.95 (-2.94,2.94) {0.25,1.75}
Total: 15832 W: 890 L: 776 D: 14166
Ptnml(0-2): 17, 669, 6429, 785, 16
https://tests.stockfishchess.org/tests/view/5f3c46a0a95672ddd56c632a
closes https://github.com/official-stockfish/Stockfish/pull/3028
see also https://github.com/official-stockfish/Stockfish/pull/3020
Bench: 4348811
2020-08-20 21:12:37 +02:00
581b92e4a7
Remove last captures extension
...
STC https://tests.stockfishchess.org/tests/view/5f395657e98b6c64b3df41dd
LLR: 2.95 (-2.94,2.94) {-1.50,0.50}
Total: 144664 W: 15426 L: 15537 D: 113701
Ptnml(0-2): 612, 11341, 48537, 11230, 612
LTC https://tests.stockfishchess.org/tests/view/5f3a2ec7b38d442594aabdd7
LLR: 2.96 (-2.94,2.94) {-1.50,0.50}
Total: 22728 W: 1161 L: 1146 D: 20421
Ptnml(0-2): 21, 960, 9388, 973, 22
closes https://github.com/official-stockfish/Stockfish/pull/3020
bench: 3832662
2020-08-18 08:33:42 +02:00
1c0b7bdf4f
Remove history bonus from Eval
...
STC:
LLR: 2.92 (-2.94,2.94) {-1.50,0.50}
Total: 26776 W: 2787 L: 2725 D: 21264
https://tests.stockfishchess.org/tests/view/5f39d6beb38d442594aabd9b
LTC:
LLR: 2.93 (-2.94,2.94) {-1.50,0.50}
Total: 12968 W: 635 L: 608 D: 11725
https://tests.stockfishchess.org/tests/view/5f39decfb38d442594aabda7
closes https://github.com/official-stockfish/Stockfish/pull/3019
Bench: 4335100
2020-08-18 08:29:42 +02:00
0e17a89e4d
Simplify away the passed pawn extension
...
STC https://tests.stockfishchess.org/tests/view/5f3955f0e98b6c64b3df41d7
LLR: 2.96 (-2.94,2.94) {-1.50,0.50}
Total: 31992 W: 3611 L: 3548 D: 24833
Ptnml(0-2): 174, 2658, 10273, 2713, 178
LTC https://tests.stockfishchess.org/tests/view/5f399e41e98b6c64b3df4210
LLR: 3.01 (-2.94,2.94) {-1.50,0.50}
Total: 29568 W: 1488 L: 1480 D: 26600
Ptnml(0-2): 40, 1272, 12142, 1300, 30
closes https://github.com/official-stockfish/Stockfish/pull/3017
bench: 3844671
-----
Recommended net: https://tests.stockfishchess.org/api/nn/nn-82215d0fd0df.nnue
2020-08-17 12:27:35 +02:00
430467db1c
Create a UCI Training option
...
Creates a UCI Training option and uses it to disable LMP on PV nodes.
2020-08-13 08:32:10 +09:00
44a54b63f1
Don't allow LMP on PvNodes
...
I mentioned this a while back in discord, but nothing seems to have ever come from it. Anyway, to the best of my knowledge most current training data gen is being done at relatively low fixed depths. With this in mind, the change to not allow LMP in PvNodes should result in a fairly significant increase in strength and reliability of the PV.
2020-08-13 08:32:10 +09:00
Tomasz Sobczyk