Stockfish Development Versions are build automatically if there are changes on the master branch in the git repository (https://github.com/official-stockfish/Stockfish). Use it at your own risk.
They are compiled with gcc 11.2/mingw 10 on Ubuntu 22.04.

Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: protonspring
Date: Sun Mar 10 10:53:39 2019 +0100
Timestamp: 1552211619

Remove popcount16() (#2038)

This is a non-functional simplification / code-style change.

This popcount16 method does nothing but initialize the PopCnt16 arrays.

This can be done in a single bitset line, which is less lines and more clear. Performance for this code is moot.

No functional change.
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: xoto10
Date: Sun Mar 10 10:47:42 2019 +0100
Timestamp: 1552211262

Simplify failedLow away #1986

FailedLow doesn't seem to add any value so remove it.

STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 43915 W: 9682 L: 9604 D: 24629 Elo +0.62
http://tests.stockfishchess.org/tests/view/5c5339770ebc592fc7baef74

LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 58515 W: 9670 L: 9609 D: 39236 Elo +0.36
http://tests.stockfishchess.org/tests/view/5c53cc840ebc592fc7baf6c1

Ideas for further work:

Tune the values in the revised fallingEval calculation
Consider adding a term using delta, e.g. c * (delta - 20) as an indicator of eval instability

Bench: 3318033
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Marco Costalba
Date: Sat Mar 9 13:28:11 2019 +0100
Timestamp: 1552134491

Revert "Allowing singular extension in mate positions"

It was causing an assert: value > -VALUE_INFINITE
under some conditions.

See https://github.com/official-stockfish/Stockfish/issues/2036

Bench: 3318033
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: protonspring
Date: Tue Mar 5 20:48:29 2019 +0100
Timestamp: 1551815309

Remove FutilityMoveCounts array. (#2024)

This is a functional simplification that removes the FutilityMoveCounts array with a simple equation using only ints.

LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 14175 W: 3123 L: 2987 D: 8065 Elo +3.33

LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 9900 W: 1735 L: 1597 D: 6568 Elo +4.84

Bench: 3380343

see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: protonspring
Date: Tue Mar 5 19:02:04 2019 +0100
Timestamp: 1551808924

Shrink Reductions[] array to one dimension

This is a non-functional patch which shrinks the reductions array.
This saves about 8Kb of memory.

The only slow part of master's reductions array is the calculation
of the log values, so using a separate array for those values and
calculating the rest real-time appears to be just as fast as master.

STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 63245 W: 13906 L: 13866 D: 35473 Elo +0.22
http://tests.stockfishchess.org/tests/view/5c7b571f0ebc5925cffdc104

No funcional change.
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: SFisGOD
Date: Tue Mar 5 14:54:25 2019 +0100
Timestamp: 1551794065

Pawn value tweak

STC:
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 47166 W: 10664 L: 10311 D: 26191 Elo +2.60
http://tests.stockfishchess.org/tests/view/5c7dfc370ebc5925cffdf830

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 35439 W: 6034 L: 5767 D: 23638 Elo +2.62
http://tests.stockfishchess.org/tests/view/5c7e41020ebc5925cffdfe9b

Bench: 3470519
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Marco Costalba
Date: Tue Mar 5 11:13:02 2019 +0100
Timestamp: 1551780782

Assorted trivial cleanups 2/2019

No functional change.
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Vizvezdenec
Date: Tue Mar 5 11:10:10 2019 +0100
Timestamp: 1551780610

Add continuation history 5

Original patch passed STC:
http://tests.stockfishchess.org/tests/view/5c7439ff0ebc5925cffd3e64
LLR: 2.95 (-2.94,2.94) [0.50,4.50]
Total: 26348 W: 5926 L: 5632 D: 14790 Elo +3.88

and LTC:
http://tests.stockfishchess.org/tests/view/5c745a8b0ebc5925cffd41a8
LLR: 2.95 (-2.94,2.94) [0.00,3.50]
Total: 198411 W: 33238 L: 32510 D: 132663 Elo +1.27

But had undefined behavior.
After fixing (thx to @vondele )

passed LTC:
http://tests.stockfishchess.org/tests/view/5c763c7c0ebc5925cffd5de2
LLR: 2.95 (-2.94,2.94) [0.00,3.50]
Total: 112253 W: 18711 L: 18225 D: 75317 Elo +1.50

bench 3049229
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Joost VandeVondele
Date: Wed Feb 27 13:36:48 2019 +0100
Timestamp: 1551271008

Remove skipQuiets with mate fix. (#2021)

This removes the skipQuiets variable, as was done in an earlier round by
@protonspring, but fixes an oversight which led to wrong mate
announcements. Quiets can only be pruned when there is no mate score, so
set moveCountPruning at the right spot.

tested as a fix at STC:
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 66321 W: 14690 L: 14657 D: 36974 Elo +0.17
http://tests.stockfishchess.org/tests/view/5c74f3170ebc5925cffd4b3c

and as the full patch at LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 25903 W: 4341 L: 4228 D: 17334 Elo +1.52
http://tests.stockfishchess.org/tests/view/5c7540030ebc5925cffd506f

Bench: 3292342
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Marco Costalba
Date: Tue Feb 26 13:23:09 2019 +0100
Timestamp: 1551183789

Revert "Remove skipQuiets variable in search()"

This reverts commit 76d2f5b94a0df20d84ccf922bd1c0fcf1c779090.

Due to a bug, see https://github.com/official-stockfish/Stockfish/issues/2019

Bench: 3516616
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Kurtbusch
Date: Thu Feb 21 19:53:03 2019 +0100
Timestamp: 1550775183

Add KNNvKP Endgame Heuristic

This is a somewhat different patch. It fixes blindspots for
two knights vs pawn endgame.

With local testing starting from random KNNvKP positions where the
pawn has not advanced beyond the 4th rank (thanks @protonspring !)
at 15+0.15 (4 cores), this went +105=868-27 against master. All except
two losses were won in reverse.

The heuristic is simple but effective - the strategy in these endgames
is to push the opposing king to the corner, then move the knight that's
blocking the pawn in for the checkmate while the pawn is free to move
and prevents stalemate. This patch gives SF the little boost it needs
to search the relevant king-cornering mating lines.

See the discussion in pull request 1939 for some more good results for
this test in independant tests:
https://github.com/official-stockfish/Stockfish/pull/1939

Bench: 3310239
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: MJZ1977
Date: Thu Feb 21 19:36:48 2019 +0100
Timestamp: 1550774208

ProbCutCount limit dependancy to cutNode

Use the ProbCutCount limit `2 + 2 * cutNode` instead of constant 3.

STC
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 61812 W: 13599 L: 13459 D: 34754 Elo +0.79
http://tests.stockfishchess.org/tests/view/5c6d19240ebc5925cffca07a

LTC
LLR: 2.96 (-2.94,2.94) [0.00,3.50]
Total: 27549 W: 4614 L: 4363 D: 18572 Elo +3.17
http://tests.stockfishchess.org/tests/view/5c6d45c10ebc5925cffca7a6

Closes https://github.com/official-stockfish/Stockfish/pull/2015

Bench: 3368889
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: protonspring
Date: Thu Feb 21 19:24:02 2019 +0100
Timestamp: 1550773442

Remove PvNode dimension from Reductions array

This is a functional simplification: if we simply subtract one to Reductions[]
when PvNode is set, we can remove this dimension of the multidimensional array.
I think this saves about 8K of memory.

STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 10118 W: 2282 L: 2138 D: 5698 Elo +4.95
http://tests.stockfishchess.org/tests/view/5c6332b60ebc5925cffbdfed

LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 70765 W: 11617 L: 11575 D: 47573 Elo +0.21
http://tests.stockfishchess.org/tests/view/5c63379e0ebc5925cffbe0de

Closes https://github.com/official-stockfish/Stockfish/pull/2010

Bench 3261078
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: protonspring
Date: Thu Feb 21 19:18:02 2019 +0100
Timestamp: 1550773082

Remove skipQuiets variable in search()

This is a functional simplification. The moveCountPruning variable and the
skipQuiets variable are similar enough in function that they can be combined.
This removes the skipQuiets variable in search.

STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 23278 W: 5210 L: 5091 D: 12977 Elo +1.78
http://tests.stockfishchess.org/tests/view/5c65dc490ebc5925cffc12e9

LTC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 77107 W: 12792 L: 12761 D: 51554 Elo +0.14
http://tests.stockfishchess.org/tests/view/5c65e4360ebc5925cffc1490

Closes https://github.com/official-stockfish/Stockfish/pull/2011

bench 3640330
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: protonspring
Date: Fri Feb 8 10:31:28 2019 +0100
Timestamp: 1549618288

Change outposts to single value #1946

This is a functional simplification of the Outposts array
moving it to a single value. This is a duplicate PR because
I couldn't figure out how to fix the original one.

The idea is from @31m059 with formatting recommendations by @snicolet.

See #1940 for additional information.

STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 23933 W: 5279 L: 5162 D: 13492 Elo +1.70
http://tests.stockfishchess.org/tests/view/5c3575800ebc596a450c5ecb

LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 41718 W: 6919 L: 6831 D: 27968 Elo +0.73
http://tests.stockfishchess.org/tests/view/5c358c440ebc596a450c6117

bench 3783543
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Marco Costalba
Date: Fri Feb 8 10:20:43 2019 +0100
Timestamp: 1549617643

Assorted trivial cleanups 1/2019

To address #1862

No functional change.
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Marco Costalba
Date: Fri Feb 8 10:19:00 2019 +0100
Timestamp: 1549617540

Log message of: Less king danger...

The commit:

Less king danger if we have a knight near by to defend it

went in withouth proper commit message, here below we add it:

STC: LLR: 2.95 (-2.94,2.94) [0.50,4.50]
Total: 113106 W: 25087 L: 24367 D: 63652 Elo +2.21
http://tests.stockfishchess.org/tests/view/5c5517540ebc592fc7bb0eb4

LTC: LLR: 2.95 (-2.94,2.94) [0.00,3.50]
Total: 96669 W: 16318 L: 15872 D: 64479 Elo +1.60
http://tests.stockfishchess.org/tests/view/5c55352b0ebc592fc7bb11c8

bench: 3653942
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: protonspring
Date: Fri Feb 8 09:54:38 2019 +0100
Timestamp: 1549616078

Remove Some Bitboard Arrays (#1963)

This is non-functional. These 5 arrays are simple enough to calculate real-time and maintaining an array for them does not help. Decreases the memory footprint.

This seems a tiny bit slower on my machine, but passed STC well enough. Could someone verify speed?

STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 44745 W: 9780 L: 9704 D: 25261 Elo +0.59
http://tests.stockfishchess.org/tests/view/5c47aa2d0ebc5902bca13fc4

The slowdown is minimal even in 32 bit case (thanks to @mstembera for testing):

Compiled using make build ARCH=x86-32 CXX=i686-w64-mingw32-c++ and benched
This patch only:

```
Results for 40 tests for each version:

Base Test Diff
Mean 1455204 1450033 5171
StDev 49452 34533 59621

p-value: 0.465
speedup: -0.004
```

No functional change.
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: mstembera
Date: Sun Feb 3 14:16:34 2019 +0100
Timestamp: 1549199794

Less king danger if we have a knight near by to defend it. (#1987)

bench: 3653942
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Miguel Lahoz
Date: Fri Feb 1 16:27:21 2019 +0100
Timestamp: 1549034841

Extend discovered checks regardless of SEE

A simple idea, but it makes sense: in current master the search is extended
for checks that are considered somewhat safe, and for for this we use the
static exchange evaluation which only considers the `to_sq` of a move.
This is not reliable for discovered checks, where another piece is giving
the check and is arguably a more dangerous type of check. Thus, if the check
is a discovered check, the result of SEE is not relevant and can be ignored.

STC:
LLR: 2.96 (-2.94,2.94) [0.50,4.50]
Total: 29370 W: 6583 L: 6274 D: 16513 Elo +3.66
http://tests.stockfishchess.org/tests/view/5c5062950ebc593af5d4d9b5

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,3.50]
Total: 227341 W: 37972 L: 37165 D: 152204 Elo +1.23
http://tests.stockfishchess.org/tests/view/5c5094fb0ebc593af5d4dc2c

Bench: 3611854
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Stéphane Nicolet
Date: Fri Feb 1 15:34:46 2019 +0100
Timestamp: 1549031686

Tweak tropism weight in king danger

There was a simplification attempt last week for the tropism
term in king danger, which passed STC but failed LTC. This
was an indirect sign that maybe the tropism factor was sightly
untuned in current master, so we tried to change it from 1/4
to 5/16.

STC:
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 28098 W: 6264 L: 5990 D: 15844 Elo +3.39
http://tests.stockfishchess.org/tests/view/5c518db60ebc593af5d4e306

LTC:
LLR: 2.95 (-2.94,2.94) [0.00,3.50]
Total: 103709 W: 17387 L: 16923 D: 69399 Elo +1.55
http://tests.stockfishchess.org/tests/view/5c52a5510ebc592fc7baea8b

Bench: 4016000
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Vizvezdenec
Date: Fri Feb 1 12:43:33 2019 +0100
Timestamp: 1549021413

More precise checks evaluation in king danger

Remove overlapping safe checks from kingdanger:
- rook and queen checks from the same square: rook check is preferred
- bishop and queen checks form the same square: queen check is preferred

Increase bishop and rook check values as a compensation.

STC
LLR: 2.95 (-2.94,2.94) [0.50,4.50]
Total: 27480 W: 6111 L: 5813 D: 15556 Elo +3.77
http://tests.stockfishchess.org/tests/view/5c521d050ebc593af5d4e66a

LTC
LLR: 2.95 (-2.94,2.94) [0.00,3.50]
Total: 78500 W: 13145 L: 12752 D: 52603 Elo +1.74
http://tests.stockfishchess.org/tests/view/5c52b9460ebc592fc7baecc5

Closes https://github.com/official-stockfish/Stockfish/pull/1983

------------------------------------------

I have quite a few ideas of how to improve this patch.

- actually rethinking it now it will maybe be useful to discount
queen/bishop checks if there is only one square that they can
give check from and it's "occupied" by more valuable check. Right
now count of this squares does not really matter.

- maybe some small extra bonus can be given for overlapping checks.

- some ideas about using popcount() on safechecks can be retried.

- tune this safecheck values since they were more or less randomly handcrafted in this patch.

Bench: 3216489
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: protonspring
Date: Thu Jan 31 15:21:24 2019 +0100
Timestamp: 1548944484

Simplify Stat Score bonus

This is a functional simplification of this statScore bonus.
There seems to be little risk of regression with this one.

STC
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 26829 W: 5892 L: 5781 D: 15156 Elo +1.44
http://tests.stockfishchess.org/tests/view/5c5086bb0ebc593af5d4db75

LTC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 28232 W: 4684 L: 4575 D: 18973 Elo +1.34
http://tests.stockfishchess.org/tests/view/5c50d7690ebc593af5d4dec9

Closes https://github.com/official-stockfish/Stockfish/pull/1979

Bench: 4001014
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: DU-jdto
Date: Tue Jan 29 17:40:00 2019 +0100
Timestamp: 1548780000

Don't update pvHit after IID

This patch removes line 875 of search.cpp, which was updating pvHit after IID.
Bench testing at depth 22 shows that line 875 of search.cpp never changes the
value of pvHit at NonPV nodes, while at PV nodes it often changes the value
from true to false (and never the reverse). This is because the definition of
pvHit at line 642 is :

```
pvHit = (ttHit && tte->pv_hit()) || (PvNode && depth > 4 * ONE_PLY);
```

while the assignment after IID omits the ` (PvNode && depth > 4 * ONE_PLY) `
condition. As such, unlike the other two post-IID tte reads, this line of code
does not make SF's state more consistent, but rather introduces an inconsistency
in the definition of pvHit. Indeed, changing line 875 read

```
pvHit = (ttHit && tte->pv_hit()) || (PvNode && depth > 4 * ONE_PLY);
```

to match line 642 is functionally equivalent to removing the line entirely, as
this patch does.

STC
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 62756 W: 13787 L: 13746 D: 35223 Elo +0.23
http://tests.stockfishchess.org/tests/view/5c446c850ebc5902bb5d4b75

LTC
LLR: 3.19 (-2.94,2.94) [-3.00,1.00]
Total: 61900 W: 10179 L: 10111 D: 41610 Elo +0.38
http://tests.stockfishchess.org/tests/view/5c45bf610ebc5902bb5d5d62

Bench: 3796134
see source
Windows x64 for Haswell CPUs
Windows x64 for modern computers
Windows x64
Windows 32
Linux x64 for Haswell CPUs
Linux x64 for modern computers
Linux x64
Author: Miguel Lahoz
Date: Tue Jan 29 17:32:41 2019 +0100
Timestamp: 1548779561

Change pinning logic in Static Exchange Evaluation (SEE)

This changes 2 parts with regards to static exchange evaluation.

Currently, we do not allow pinned pieces to recapture if *all* opponent
pinners are still in their starting squares. This changes that to having
a less strict requirement, checking if *any* pinners are still in their
starting square. This makes our SEE give more respect to the pinning
side with regards to exchanges, which makes sense because it helps our
search explore more tactical options.

Furthermore, we change the logic for saving pinners into our state
variable when computing slider_blockers. We will include double pinners,
where two sliders may be looking at the same blocker, a similar concept
to our mobility calculation for sliders in our evaluation section.
Interestingly, I think SEE is the only place where the pinners bitboard
is actually used, so as far as I know there are no other side effects
to this change.

An example and some insights:

White Bf2, Kg1
Black Qe3, Bc5

The move Qg3 will be given the correct value of 0. (Previously < 0)
The move Qd4 will be incorrectly given a value of 0. (Previously < 0)

It seems the tradeoff in search is worth it. Qd4 will likely be pruned
soon by something like probcut anyway, while Qg3 could help us spot
tactics at an earlier depth.

STC:
LLR: 2.96 (-2.94,2.94) [0.50,4.50]
Total: 62162 W: 13879 L: 13408 D: 34875 Elo +2.63
http://tests.stockfishchess.org/tests/view/5c4ba1a70ebc593af5d49c55

LTC: (Thanks to @alayant)
LLR: 3.40 (-2.94,2.94) [0.00,3.50]
Total: 140285 W: 23416 L: 22825 D: 94044 Elo +1.46
http://tests.stockfishchess.org/tests/view/5c4bcfba0ebc593af5d49ea8

Bench: 3937213
see source

< prev page next page >