Stockfish и его производные

[Харлампий]

Stockfish 14
Stockfish 14 для Android
https://ufile.io/f/8ardx
https://pixeldrain.com/l/ndVpnHXc#item=1

Архив предыдущих версий: https://www.dropbox.com/sh/75gzfgu7q...UfHbBr4r_RwhGa

Официальный сайт стокфиш

Обновления тестовых версий стокфиша: http://abrok.eu/stockfish/

Сборки Stockfish и CorChess от I. Ivec

Развитие движка http://tests.stockfishchess.org/tests

Stockfish_NNUE https://github.com/nodchip/Stockfish

Сети: https://www.comp.nus.edu.sg/~sergio-v/nnue/

AsmFish

Сборки от Maratka

Компиляции движков от ChessMan

Сборка Stockfish от Otodranik

Sugar от Marco Zerbinati

Shark и Orka

О Stockfish

Цитата:

Авторы движка
Двигатель Stockfish был разработан Tord Romstad, Marco Costalba и Joona Kiiski. В настоящее время он развивается и поддерживается сообществом Stockfish.

Авторы приложений
Stockfish для Mac был собран Daylen Yang. Stockfish для iOS был собран Tord Romstad.

История
Проект Stockfish начался с движка Glaurung с открытым исходным кодом, созданного Tord Romstad. В ноябре 2008 года Марко Косталба разветвил код Glaurung 2.1 и представил Stockfish 1.0. Торд и Джоона Кииски присоединились к проекту «Стилфиш», и проект «Гларунг» постепенно исчез. Тем временем, Stockfish быстро поднялся, чтобы стать самым сильным шахматным движком с открытым исходным кодом, с частыми обновлениями каждые несколько месяцев. Сегодня он остается одним из самых сильных двигателей в мире.

О GPL
Stockfish является бесплатным и распространяется в соответствии с GNU General Public License Version 3 (GPLv3). По сути, это означает, что вы можете делать практически все, что хотите, с программой, включая ее распространение среди своих друзей, делая ее доступной для загрузки с вашего веб-сайта, продавая ее (либо самостоятельно, либо как часть некоторого более крупного программного пакета ) или использовать его в качестве отправной точки для собственного программного проекта.

Единственное реальное ограничение заключается в том, что всякий раз, когда вы каким-либо образом распространяете Stockfish, вы всегда должны включать полный исходный код или ссылку на исходный код. Если вы внесете какие-либо изменения в исходный код, эти изменения также должны быть доступны в GPL.

Для получения полной информации прочитайте GPL.

[свернуть]

[Харлампий]

Sting SF 9.9: http://www.mediafire.com/file/ii74se...ting-sf-99.zip

[Кевин]

Author: DU-jdto
Date: Wed Mar 21 01:35:05 2018 +0100
Timestamp: 1521592505

Fix a few minor code style inconsistencies

No functional change.

http://abrok.eu/stockfish/

[Кевин]

Author: Joost VandeVondele
Date: Wed Mar 21 01:48:47 2018 +0100
Timestamp: 1521593327

Join refutation stages in the movepicker

Unifies a bit further the three refuation stages in the MovePicker
class. Also treat the skipping of TT move now always via select_move(),
as discussed in pull request #1454.

Passed STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 16608 W: 3461 L: 3331 D: 9816
http://tests.stockfishchess.org/test...bc59029fb6f6c3

Closes https://github.com/official-stockfis...fish/pull/1502

No functional change.


http://abrok.eu/stockfish/

[Александр Руденко]

Author: Chess13234
Date: Sat Mar 24 08:59:48 2018 +0100
Timestamp: 1521878388

IID margin reduced to 128

Change the internal iterative deepening margin from +256 to +128.

Passed STC:
LLR: 3.26 (-2.94,2.94) [0.00,5.00]
Total: 22006 W: 4646 L: 4400 D: 12960
http://tests.stockfishchess.org/test...bc590295d8847d

Passed LTC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 46665 W: 7187 L: 6898 D: 32580
http://tests.stockfishchess.org/test...bc590295d884b4

Possible work in this area after this change:

• Reduce even more the margin?
• Try to do IID for smaller depths too?

Bench: 5851102

http://abrok.eu/stockfish/

[Кевин]

http://chess.ultimaiq.net/stockfish.html
Stockfish development builds compiled by I. Ivec and L. Ljubičić
March 24, 2018

[ChessMan]

Кто сомневался, что динамический контемпт плох для анализа, вот почитайте: https://groups.google.com/forum/m/#!...U/_CX-tzeeCAAJ


Регрессия в MultiPV.

[Кевин]

Цитата: Сообщение от ChessMan

Кто сомневался, что динамический контемпт плох для анализа

Продолжу: тот мало разбирается в анализе

Впрочем, я уже писал, что новый Сток не так плох (в одну линию) в некоторых проверенных мною позициях, а вот если с мультиком проблемы - это действительно не айс.

[Кевин]

Author: Jerry Donald Watson
Date: Mon Mar 26 09:28:37 2018 +0200
Timestamp: 1522049317

Make kingRing always 8 squares

Make kingRing always eight squares, extending the bitboard to the
F file if the king is on the H file, and to the C file if the king
is on the A file. This may deal with cases where Stockfish (like
many other engines) would shift the king around on the back rank
like g1h1, not because there is some imminent threat, but because
it makes king safety look a little better just because the king ring
had a smaller area.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 34000 W: 7167 L: 6877 D: 19956
http://tests.stockfishchess.org/test...bc5902932cbe64

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 22574 W: 3576 L: 3370 D: 15628
http://tests.stockfishchess.org/test...bc5902932cbe72

How to continue from there?

This patch probably makes it easier to tune the king safety evaluation,
because the new regularity of the king ring size will make the king
safety function more continuous.

Closes https://github.com/official-stockfis...fish/pull/1512

Bench: 5934103


http://abrok.eu/stockfish/

[Валекс]

Цитата: Сообщение от ChessMan

Кто сомневался, что динамический контемпт плох для анализа, вот почитайте: https://groups.google.com/forum/m/#!...U/_CX-tzeeCAAJ


Регрессия в MultiPV.

Поясните пожалуйста, как это влияет на анализе в ИДЕЕ?

[ChessMan]

Цитата: Сообщение от Валекс

Поясните пожалуйста, как это влияет на анализе в ИДЕЕ?

IDeA не использую, так что понятия не имею.

[Кевин]

http://chess.ultimaiq.net/stockfish.html
Stockfish development builds compiled by I. Ivec and L. Ljubičić
March 26, 2018

[Харлампий]

Аж шесть обновлений сегодня: http://abrok.eu/stockfish/

[svoitsl]

Цитата: Сообщение от Харлампий

Аж шесть обновлений сегодня:

наверное к суперфиналу TCEC 11 готовятся....

[Кевин]

Author: Jerry Donald Watson
Date: Wed Mar 28 00:31:47 2018 +0200
Timestamp: 1522189907

Change reduction/improving interaction

Adjust criterion for applying extra reduction if not improving.
We now add an extra ply of reduction if r > 1.0, instead of the
previous condition Reductions[NonPV][imp][d][mc] >= 2.

Why does this work? Previously, reductions when not improving had
a discontinuity as the depth and/or move count increases due to the
Reductions[NonPV][imp][d][mc] >= 2 condition. Hence, values of r
such that 0.5 < r < 1.5 would be mapped to a reduction of 1, while
1.5 < r < 2.5 would be mapped to a reduction of 3. This patch allows
values of r satisfying 1.0 < r < 1.5 to be mapped to a reduction of 2,
making the reduction formula more continuous.

STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 35908 W: 7382 L: 7087 D: 21439
http://tests.stockfishchess.org/test...bc5902a4743e8f

LTC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 23087 W: 3584 L: 3378 D: 16125
http://tests.stockfishchess.org/test...bc5902a4743ea9

Ideas for future work:
- We could look at retuning the LMR formula.
- We could look at adjusting the reductions in PV nodes if not improving.

Bench: 5326261


http://abrok.eu/stockfish/

[Кевин]

Цитата: Сообщение от Харлампий

Аж шесть обновлений сегодня: http://abrok.eu/stockfish/

Причём в одном из них - исправление взаимодействия мульти-пв и динамического контемпта. Молодцы

[Валекс]

Цитата: Сообщение от Кевин

Причём в одном из них - исправление взаимодействия мульти-пв и динамического контемпта. Молодцы

Исправление, как я понял, у Jerry Donald Watson?

[otodranik]

Цитата: Сообщение от Валекс

Исправление, как я понял, у Jerry Donald Watson?

Все последующие компиляции учитывают все правки предыдущих.

[ChessMan]

Цитата: Сообщение от Валекс

Исправление, как я понял, у Jerry Donald Watson?

Нет. У WOnder93
https://github.com/official-stockfis...f25c2a9b60de8f

======
И да, все последующие патчи включают эту правку.

[Кевин]

Цитата: Сообщение от Валекс

Исправление, как я понял, у Jerry Donald Watson?

Ondrej Mosnáček автор исправления.

[Кевин]

Регресс-тест Стока:


27-03-18 sn master diff
ELO: 13.77 +-1.7 (95%) LOS: 100.0%
Total: 40000 W: 5924 L: 4339 D: 29737
40000 @ 60+0.6 th 1 Regression/progression test after "Make kingRing always 8 squares" against SF9.


Прошлый тест:
14-03-18 sn master diff
ELO: 13.03 +-1.7 (95%) LOS: 100.0%
Total: 40000 W: 5963 L: 4463 D: 29574
40000 @ 60+0.6 th 1 Regression/progression test after "Use intrinsics only for LSB/MSB" against SF9.

[Кевин]

Author: Ondrej Mosnáček
Date: Fri Mar 30 10:48:57 2018 +0200
Timestamp: 1522399737

Use per-thread dynamic contempt

We now use per-thread dynamic contempt. This patch has the following
effects:

Развернуть для просмотра

* for Threads=1: **non-functional**
* for Threads>1:
* with MultiPV=1: **no regression, little to no ELO gain**
* with MultiPV>1: **clear improvement over master**

First, I tried testing at standard MultiPV=1 play with [0,5] bounds.
This yielded 2 yellow and 1 red test:

5+0.05, Threads=5:
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 82689 W: 16439 L: 16190 D: 50060
http://tests.stockfishchess.org/tests/view/5aa93a5a0ebc5902952892e6

5+0.05, Threads=8:
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 27164 W: 4974 L: 4983 D: 17207
http://tests.stockfishchess.org/tests/view/5ab2639b0ebc5902a6fbefd5

5+0.5, Threads=16:
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 41396 W: 7127 L: 7082 D: 27187
http://tests.stockfishchess.org/tests/view/5ab124220ebc59029516cb62

Then, I tested with Skill Level=17 (implicitly MutliPV=4), showing
a clear improvement:

5+0.05, Threads=5:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 3498 W: 1316 L: 1135 D: 1047
http://tests.stockfishchess.org/tests/view/5ab4b6580ebc5902932aeca2

Next, I tested the patch with MultiPV=1 again, this time checking for
non-regression ([-3, 1]):

5+0.5, Threads=5:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 65575 W: 12786 L: 12745 D: 40044
http://tests.stockfishchess.org/tests/view/5ab4e8500ebc5902932aecb3

Finally, I ran some tests with fixed number of games, checking if
reverting dynamic contempt gains more elo with Skill Level=17 (i.e.
MultiPV) than applying the "prevScore" fix and this patch. These tests
showed, that this patch gains 15 ELO when playing with Skill Level=17:

5+0.05, Threads=3, "revert dynamic contempt" vs. "WITHOUT this patch":
ELO: -11.43 +-4.1 (95%) LOS: 0.0%
Total: 20000 W: 7085 L: 7743 D: 5172
http://tests.stockfishchess.org/tests/view/5ab636450ebc590295d88536

5+0.05, Threads=3, "revert dynamic contempt" vs. "WITH this patch":
ELO: -26.42 +-4.1 (95%) LOS: 0.0%
Total: 20000 W: 6661 L: 8179 D: 5160
http://tests.stockfishchess.org/tests/view/5ab62e680ebc590295d88524

---
***FAQ***

**Why should this be commited?**
I believe that the gain for multi-thread MultiPV search is a sufficient
justification for this otherwise neutral change. I also believe this
implementation of dynamic contempt is more logical, although this may
be just my opinion.

**Why is per-thread contempt better at MultiPV?**
A likely explanation for the gain in MultiPV mode is that during
search each thread independently switches between rootMoves and via
the shared contempt score skews each other's evaluation.

**Why were the tests done with Skill Level=17?**
This was originally suggested by @Hanamuke and the idea is that with
Skill Level Stockfish sometimes plays also moves it thinks are slightly
sub-optimal and thus the quality of all moves offered by the MultiPV
search is checked by the test.

**Why are the ELO differences so huge?**
This is most likely because of the nature of Skill Level mode --
since it slower and weaker than normal mode, bugs in evaluation have
much greater effect.

---

Closes https://github.com/official-stockfish/Stockfish/pull/1515.

[свернуть]

No functional change -- in single thread mode



Author: Joost VandeVondele
Date: Fri Mar 30 10:23:48 2018 +0200
Timestamp: 1522398228

Integrate syzygy in automated testing (v2).

Extends valgrind/sanitizer testing to cover syzygy code.

The script downloads 4 man syzygy as needed. The time needed for the
additional testing is small (in fact hard to see a difference compared
to the large fluctuations in testing time in travis).

Possible follow-ups:

* include more TB sensitive positions in bench.
* include the test script of recent commit "Refactor tbprobe.cpp".
* verify unchanged bench with TB (with a long run).
* make the TB part of the continuation integration tests optional.

Closes https://github.com/official-stockfish/Stockfish/pull/1518
and https://github.com/official-stockfish/Stockfish/pull/1490

No functional change.


http://abrok.eu/stockfish/

[Кевин]

http://chess.ultimaiq.net/stockfish.html
Stockfish development builds compiled by I. Ivec and L. Ljubičić
March 30, 2018

[Кевин]

Author: Alain SAVARD
Date: Sat Mar 31 01:16:51 2018 +0200
Timestamp: 1522451811

Candidate Passed Pawn

Include some not fully supported levers in the (candidate) passed pawns
bitboard, if otherwise unblocked. Maybe levers are usually very short
lived, and some inaccuracy in the lever balance for the definition of
candidate passed pawns just triggers a deeper search.

Here is a example of a case where the patch has an effect on the definition
of candidate passers: White c5/e5 pawns, against Black d6 pawn. Let's say
we want to test if e5 is a candidate passer. The previous master looks
only at files d, e and f (which is already very good) and reject e5 as
a candidate. However, the lever d6 is challenged by 2 pawns, so it should
not fully count. Indirectly, this patch will view such case (and a few more)
to be scored as candidates.

STC
http://tests.stockfishchess.org/test...bc5902926cf1e1
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 16492 W: 3419 L: 3198 D: 9875

LTC
http://tests.stockfishchess.org/test...bc5902926cf1e6
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 21156 W: 3201 L: 2990 D: 14965

This was inspired by this test of Jerry Donald Watson, except the case of
zero supporting pawns against two levers is excluded, and it seems that
not excluding that case is bad, while excluding is it beneficial. See the
following tests on fishtest:

https://github.com/official-stockfis...fish/pull/1519
http://tests.stockfishchess.org/test...bc5902926cf1dd
http://tests.stockfishchess.org/test...bc5902926cf1e4

Closes https://github.com/official-stockfis...fish/pull/1521

Bench: 5568461

----

Comments by Jerry Donald Watson:

Развернуть для просмотра

> My thinking as to why this works:
>
> The evaluation is either called in an interior node or in the qsearch.
> The calls at the end of the qsearch are the more important as they
> ultimately determine the scoring of each move, whereas the internal
> values are mainly used for pruning decisions with a margin. Some strong
> engines don't even call the eval at all nodes. Now the whole point of
> the qsearch is to find quiet positions where captures do not change the
> evaluation of the position with regards to the search bounds - i.e. if
> there were good captures they would be tried.* So when a candidate lever
> appears in the evaluation at the end of the qsearch, the qsearch has
> guaranteed that it cannot just be captured, or if it can, this does not
> take the score past the search bounds. Practically this may mean that
> the side with the candidate lever has the turn, or perhaps the stopping
> lever pawn is pinned, or that side is forced for other reasons to make
> some other move (e.g. d6 can only take one of the pawns in the example
> above).
>
> Hence granting the full score for only one lever defender makes some
> sense, at least, to me.
>
> IMO this is also why huge bonuses for possible captures in the evaluation
> (e.g. threat on queen and our turn), etc. don't tend to work. Such things
> are best left to the search to figure out.

[свернуть]

http://abrok.eu/stockfish/

[sovaz1997]

Новый патч добавляет 5.5 пунктов ЭЛО: http://hardy.uhasselt.be/Toga/live_e...captPrune2...&

[Кевин]

Author: Joost VandeVondele
Date: Sat Mar 31 23:10:35 2018 +0200
Timestamp: 1522530635

Refine SEE threshold for capture pruning.

eloDoc suggests that this part of search is worth ~18 Elo.
This patch refines the depth dependence of the SEE threshold.

passed STC:
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 21398 W: 4474 L: 4245 D: 12679
http://tests.stockfishchess.org/test...bc591a560aae07

passed LTC:
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 9028 W: 1439 L: 1285 D: 6304
http://tests.stockfishchess.org/test...bc591a560aae11

Closes https://github.com/official-stockfis...fish/pull/1527

Bench: 6036915


http://abrok.eu/stockfish/

[Кевин]

Author: Joost VandeVondele
Date: Sun Apr 1 03:21:45 2018 +0200
Timestamp: 1522545705

Document Elo impact of various parts of search

Развернуть для просмотра

In order to understand better the impact of various techniques used in search,
Elo estimates have been run at STC for 60000 games (statistical error ~1.8 Elo),
disabling each feature in turn. This should help future improvements and
simplifications to pick suitable targets.

The list of tests is:
step 7 : http://tests.stockfishchess.org/tests/view/5abcbb4b0ebc5902926cf1ca
step 8 : http://tests.stockfishchess.org/tests/view/5abcbb680ebc5902926cf1cc
step 9 : http://tests.stockfishchess.org/tests/view/5abcbb850ebc5902926cf1ce
step 10 : http://tests.stockfishchess.org/tests/view/5abcbbeb0ebc5902926cf1d2
step 11 : http://tests.stockfishchess.org/tests/view/5abcbbbf0ebc5902926cf1d0
step 13 : http://tests.stockfishchess.org/tests/view/5abd03680ebc5902926cf20b
step 13a: http://tests.stockfishchess.org/tests/view/5abd29660ebc5902926cf22a
step 13b: http://tests.stockfishchess.org/tests/view/5abd29820ebc5902926cf22c
step 14 : http://tests.stockfishchess.org/tests/view/5abd03860ebc5902926cf20f
step 14a: http://tests.stockfishchess.org/tests/view/5abd2b6c0ebc5902926cf230
step 14b: http://tests.stockfishchess.org/tests/view/5abd2b8d0ebc5902926cf232
step 14c: http://tests.stockfishchess.org/tests/view/5abd2bad0ebc5902926cf234
step 14d: http://tests.stockfishchess.org/tests/view/5abd2bcf0ebc5902926cf236
step 14e: http://tests.stockfishchess.org/tests/view/5abd2bf10ebc5902926cf238

This patch documents this in the code.

Note that it will be a waste to recompute these estimates often, even a couple
of [0,5] patches are unlikely to change them by more than the error bars. The
interest of the Elo annotations in the code is not in the details, but in high-
lighting trends such as razoring (2 Elo) and singular extensions (60 Elo). These
estimates should be recomputed at most once a year.

Closes https://github.com/official-stockfish/Stockfish/pull/1522

[свернуть]

No functional change.



Author: mstembera
Date: Sun Apr 1 02:32:55 2018 +0200
Timestamp: 1522542775

Some small changes

1) Use make_bitboard() in Bitboards::init()

2) Fix MSVC warning: search.h(85): warning C4244: '=': conversion from
'TimePoint' to 'int', possible loss of data.

Closes https://github.com/official-stockfish/Stockfish/pull/1524

No functional change.


Author: Stéphane Nicolet
Date: Sun Apr 1 02:19:42 2018 +0200
Timestamp: 1522541982

Simplify KBPKB endgame with opposite bishops

Развернуть для просмотра

When we reach a position with only two opposite colored bishops and
one pawn on the board, current master would give it a scale factor
of 9/64=0.14 in about one position out of 7200, and a scale factor
of 0.0 in the 7199 others. The patch gives a scale factor of 0.0 in
100% of the cases.

STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 55845 W: 11467 L: 11410 D: 32968
http://tests.stockfishchess.org/test...bc5902926cf15e

LTC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 11915 W: 1852 L: 1719 D: 8344
http://tests.stockfishchess.org/test...bc5902926cf18c

We also have exhaustive coverage analysis of this patch effect by
Alain Savard, comparing the perfect evaluation given by the Syzygy
tablebase with the heuristic play after this patch for the set of
all legal positions of the KBPKP endgame with opposite bishops, in
the comments thread for this pull request:
https://github.com/official-stockfis...fish/pull/1520

Alain's conclusion:
> According to this definition and the data, I consider this PR is
> identical to master to "solve for draw" and slightly better than
> master to solve earlier for "wins".

Note: this patch is a side effect of an ongoing effort to improve
the evaluation of positions involving a pair of opposite bishops.
See the GitHub diff of this LTC test which almost passed at sprt[0..5]
for a discussion:
http://tests.stockfishchess.org/test...bc5902932cbf93

[свернуть]

No functional change (at small bench depths)


http://abrok.eu/stockfish/

[byan_mivv]

BrainFish_180401

http://zipproth.de

[ChessMan]

Обновление асмФиш от 010418, Mohammed Li вернулся к проекту.


Скачать для Windows: https://github.com/lantonov/asmFish/...owsOS_binaries

[Кевин]

http://chess.ultimaiq.net/stockfish.html
Stockfish development builds compiled by I. Ivec and L. Ljubičić
April 1, 2018

[Кевин]

Author: Joost VandeVondele
Date: Tue Apr 3 00:13:38 2018 +0200
Timestamp: 1522707218

Simplify Singular Extension condition

Avoid defining a singly used variable, removes one condition.

passed STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 53489 W: 10814 L: 10752 D: 31923
http://tests.stockfishchess.org/test...bc590e9457cd94

Closes https://github.com/official-stockfis...fish/pull/1530

No functional change.

http://abrok.eu/stockfish/