search

amchess / ShashChess

A try to implement Alexander Shashin's theory on a Stockfish's derived chess engine
GNU General Public License v3.0
130 stars 28 forks source link

readme

Introduction

ShashChess is a free UCI chess engine derived from Stockfish family chess engines. For the evaluation function, we utilize the collaboration between Leela Chess Zero and Stockfish, for which we express our sincere gratitude. The goal is to apply Alexander Shashin theory exposed on the following book : https://www.amazon.com/Best-Play-Method-Discovering-Strongest/dp/1936277468 to improve

Also during the search, to enhance it, we use both standard and Q/Self reinforcement learning.

Terms of use

Shashchess is free, and distributed under the GNU General Public License (GPL). Essentially, this

means that you are free to do almost exactly what you want with the program, including distributing

it among your friends, making it available for download from your web site, selling it (either by

itself or as part of some bigger software package), or using it as the starting point for a software

project of your own.

The only real limitation is that whenever you distribute ShashChess in some way, you must always

include the full source code, or a pointer to where the source code can be found. If you make any

changes to the source code, these changes must also be made available under the GPL.

For full details, read the copy of the GPL found in the file named Copying.txt.

Files

This distribution of ShashChessPro consists of the following files:

Uci options

Hash Memory

Hash

Integer, Default: 16, Min: 1, Max: 131072 MB (64-bit) : 2048 MB (32-bit)

The amount of memory to use for the hash during search, specified in MB (megabytes). This number should be smaller than the amount of physical memory for your system. A modern formula to determine it is the following:

(T x S / 100) MB where T = the average move time (in seconds) S = the average node speed of your hardware A traditional formula is the following: (N x F x T) / 512 where N = logical threads number F = clock single processor frequency (MB) T = the average move time (in seconds)

Clear Hash

Button to clear the Hash Memory. If the Never Clear Hash option is enabled, this button doesn't do anything.

Threads

Integer, Default: 1, Min: 1, Max: 512 The number of threads to use during the search. This number should be set to the number of cores (physical+logical) in your CPU.

Ponder (checkbox)

Boolean, Default: True Also called "Permanent Brain" : whether or not the engine should analyze when it is the opponent's turn.

Usually not on the configuration window.

MultiPV

Integer, Default: 1, Min: 1, Max: 500 The number of alternate lines of analysis to display. Specify 1 to just get the best line. Asking for more lines slows down the search. Usually not on the configuration window.

UCI_Chess960 (checkbox)

Whether or not ShashChess should play using Chess 960 mode. Usually not on the configuration window.

Move overhead

Default 30, min 0, max 5000 In ms, the default value seems to be the best on Linux systems, but must be increased for slow GUI like Fritz. In general, on Windows system it seems a good value to be 100.

Slow mover

Default 84, min 10, max 1000 "Time usage percent": how much the engine thinks on a move. Many engines seem to move faster and the engine is behind in time clock. With lower values it plays faster, with higher values slower - of course always within the time control.

Sygyzy End Game table bases

Download at http://olympuschess.com/egtb/sbases (by Ronald De Man)

SyzygyPath

The path to the Syzygy endgame tablebases.this defines an absolute path on your computer to the tablebase files, also on multiple paths separated with a semicolon (;) character (Windows), the colon (:) character (OS X and Windows) character. The folder(s) containing the Syzygy EGTB files. If multiple folders are used, separate them by the ; (semicolon) character.

SygyzyProbeDepth

Integer, Default: 1, Min: 1, Max: 100 The probing tablebases depth (always the root position). If you don't have a SSD HD,you have to set it to maximize the depth and kn/s in infinite analysis and during a time equals to the double of that corresponding to half RAM size. Choice a test position with a few pieces on the board (from 7 to 12). For example:

Syzygy50MoveRule

Disable to let fifty-move rule draws detected by Syzygy tablebase probes count as wins or losses. This is useful for ICCF correspondence games.

SygyzyProbeLimit

Integer, Default: 6, Min: 0, Max: 6 How many pieces need to be on the board before ShashChess begins probing (even at the root). Current default, obviously, is for 6-man.

Advanced Chess Analyzer

Advanced analysis options, highly recommended for CC play

Full depth threads

Integer, Default: 0, Min: 0, Max: 512 The number of settled threads to use for a full depth brute force search. If the number is greater than threads number, all threads are for full depth brute force search.

MonteCarlo Tree Search section (experimental: thanks to original Stephan Nicolet work)

Boolean, Default: False If activated, thanks to Shashin theory, the engine will use the MonteCarlo Tree Search for Capablanca quiescent type positions and also for caos ones, in the manner specified by the following parameters. The idea is to exploit Lc0 best results in those positions types, because Lc0 uses mcts in the search.

MCTSThreads

Integer, Default: 0, Min: 0, Max: 512 The number of settled threads to use for MCTS search except the first (main) one always for alpha-beta search. In particular, if the number is greater than threads number, they will all do a montecarlo tree search, always except the first (main) for alpha-beta search.

MCTS Multi Strategy

Integer, Default: 20, Min: 0, Max: 100 Only in multi mcts mode, for tree policy.

MCTS Multi MinVisits

Integer, Default: 5, Min: 0, Max: 1000 Only in multi mcts mode, for Upper Confidence Bound.

Live Book section (thanks to Eman's author Khalid Omar for windows builds)

LiveBook Proxy Url

String, Default: "" (empty string)
The proxy URL to use for the live book. If empty, no proxy is used. The proxy should use the ChessDB REST API format.

LiveBook Proxy Diversity

Boolean, Default: False
If enabled, the engine will play a random (best) move by the proxy (query and not querybest action).

LiveBook Lichess Games

Boolean, Default: False
If enabled, the engine will use the Lichess live book by querying the Lichess API to access the game database available on the site. This option allows the engine to access a wide range of games played on Lichess to enhance its move choices.

LiveBook Lichess Masters

Boolean, Default: False
If enabled, the engine will use the Lichess live book specifically for masters' games. This allows the engine to analyze games played at a high level and utilize the best moves made by master-level players.

LiveBook Lichess Player

String, Default: "" (empty string)
The Lichess player name to use for the live book. If left empty, the engine will not query for the specific player's game data. This option is useful for studying or adapting the engine to a particular player's style.

LiveBook Lichess Player Color

String, Default: "White"
Specifies the color the engine will play as in the Lichess live book for the specified player.

LiveBook ChessDB

Boolean, Default: False
If enabled, the engine will use the ChessDB live book by querying the ChessDB API.

LiveBook Depth

Integer, Default: 255, Min: 1, Max: 255
Specifies the depth to reach using the live book in plies. The depth determines how many half-moves the engine will consider from the current position.

ChessDB Tablebase

Boolean, Default: False
If enabled, allows the engine to query the ChessDB API for Tablebase data, up to 7 pieces. This provides perfect endgame knowledge for positions with up to 7 pieces.

Lichess Tablebase

Boolean, Default: False
If enabled, allows the engine to query the Lichess API for Tablebase data, up to 7 pieces. This option also provides perfect endgame knowledge for positions with up to 7 pieces.

ChessDB Contribute

Boolean, Default: False
If enabled, allows the engine to store a move in the queue of ChessDb to be analyzed.

Full depth threads

Default 0, min 0, max 512 The number of threads doing a full depth analysis (brute force). Useful in analysis of particular hard positions to limit the strong pruning's drawbacks.

Variety (checkbox)

Default is Off: no variety. The other values are "Standard" (no elo loss: randomicity in Capablanca zone) and Psychological (randomicity in Caos zones max).

Concurrent Experience

Boolean, Default: False Set this option to true when running under CuteChess and you experiences problems with concurrency > 1 When this option is true, the saved experience file name will be modified to something like experience-64a4c665c57504a4.bin (64a4c665c57504a4 is random). Each concurrent instance of BrainLearn will have its own experience file name, however, all the concurrent instances will read "experience.bin" at start up.

Persisted learning (checkbox)

Default is Off: no learning algorithm. The other values are "Standard" and "Self", this last to activate the Q-learning, optimized for self play. Some GUIs don't write the experience file in some game's modes because the uci protocol is differently implemented

The persisted learning is based on a collection of one or more positions stored with the following format (similar to in memory Stockfish Transposition Table):

This file is loaded in an hashtable at the engine load and updated each time the engine receive quit or stop uci command. When BrainLearn starts a new game or when we have max 8 pieces on the chessboard, the learning is activated and the hash table updated each time the engine has a best score at a depth >= 4 PLIES, according to Stockfish aspiration window.

At the engine loading, there is an automatic merge to experience.exp files, if we put the other ones, based on the following convention:

<fileType><qualityIndex>.exp

where

N.B.

Because of disk access, to be effective, the learning must be made at no bullet time controls (less than 5 minutes/game).

Read only learning

Boolean, Default: False If activated, the learning file is only read.

Experience Book

Boolean, Default: False If activated, the engine will use the experience file as the book. In choosing the move to play, the engine will be based first on maximum win probability, then, on the engine's internal score, and finally, on depth. The UCI token “showexp” allows the book to display moves on a given position.

Experience Book Max Moves

Integer, Default: 100, Min: 1, Max: 100 The maximum number of moves the engine chooses from the experience book

Experience Book Min Depth

Integer, Default: 4, Min: 1, Max: 255 The min depth for the experience book

Shashin section

Default: no option settled The engine will determine dynamically the position's type starting from a "Capablanca/default positions". If one or more (mixed algorithms/positions types at the boundaries) of the seven following options are settled, it will force the initial position/algorithm understanding If, in the wdl model, we define wdl_w=Win percentage, wdl_d=Drawn percentage and Win probability=(2*wdl_w+wdl_d)/10, we have the following mapping:

Win probability range Shashin position’s type Informator symbols
[0, 6] High Petrosian -+
[7, 11] Middle-High Petrosian -+ \ -/+
[12,14] Middle Petrosian -/+
[15,20] Middle-Low Petrosian -/+ \ =/+
[21,24] Low Petrosian =/+
[24,49] Caos: Capablanca-Low Petrosian =/+ \ =
[50] Capablanca =
[51,76] Caos: Capablanca-Low Tal = \ +/=
[77,79] Low Tal +/=
[80,85] Low-Middle Tal +/= +/-
[86,88] Middle Tal +/-
[89,93] Middle-High Tal +/- \ +-
[94,100] High Tal +-

N.B. The winProbability also take into account the depth at which a move has been calculated. So, it's more effective than the cp.

Tal

Attack position/algorithm

Capablanca

Strategical algorithm (for quiescent positions)

Petrosian

Defense position/algorithm (the "reversed colors" Tal)

Acknowledgments

Stockfish community

ShashChess team

Sorry If I forgot someone.

Stockfish NNUE

Overview

Build Status Build Status

Stockfish is a free, powerful UCI chess engine derived from Glaurung 2.1. Stockfish is not a complete chess program and requires a UCI-compatible graphical user interface (GUI) (e.g. XBoard with PolyGlot, Scid, Cute Chess, eboard, Arena, Sigma Chess, Shredder, Chess Partner or Fritz) in order to be used comfortably. Read the documentation for your GUI of choice for information about how to use Stockfish with it.

The Stockfish engine features the NNUE evaluation based on efficiently updateable neural networks. The NNUE evaluation benefits from the vector intrinsics available on most CPUs (sse2, avx2, neon, or similar).

Files

This distribution of Stockfish consists of the following files:

Note: to use the NNUE evaluation, the additional data file with neural network parameters needs to be available. Normally, this file is already embedded in the binary or it can be downloaded. The filename for the default (recommended) net can be found as the default value of the EvalFile UCI option, with the format nn-[SHA256 first 12 digits].nnue (for instance, nn-c157e0a5755b.nnue). This file can be downloaded from

https://tests.stockfishchess.org/api/nn/[filename]

replacing [filename] as needed.

UCI options

Currently, Stockfish has the following UCI options:

A note on the NNUE evaluation

The NNUE evaluation computes this value with a neural network based on basic inputs (e.g. piece positions only). The network is optimized and trained on the evaluations of millions of positions at moderate search depth.

The NNUE evaluation was first introduced in shogi, and ported to Stockfish afterward. It can be evaluated efficiently on CPUs, and exploits the fact that only parts of the neural network need to be updated after a typical chess move. The nodchip repository provides additional tools to train and develop the NNUE networks.

On CPUs supporting modern vector instructions (avx2 and similar), the NNUE evaluation results in stronger playing strength, even if the nodes per second computed by the engine is somewhat lower (roughly 60% of nps is typical).

Note that the NNUE evaluation depends on the Stockfish binary and the network parameter file (see EvalFile). Not every parameter file is compatible with a given Stockfish binary. The default value of the EvalFile UCI option is the name of a network that is guaranteed to be compatible with that binary.

What to expect from Syzygybases?

If the engine is searching a position that is not in the tablebases (e.g. a position with 8 pieces), it will access the tablebases during the search. If the engine reports a very large score (typically 153.xx), this means that it has found a winning line into a tablebase position.

If the engine is given a position to search that is in the tablebases, it will use the tablebases at the beginning of the search to preselect all good moves, i.e. all moves that preserve the win or preserve the draw while taking into account the 50-move rule. It will then perform a search only on those moves. The engine will not move immediately, unless there is only a single good move. The engine likely will not report a mate score even if the position is known to be won.

It is therefore clear that this behaviour is not identical to what one might be used to with Nalimov tablebases. There are technical reasons for this difference, the main technical reason being that Nalimov tablebases use the DTM metric (distance-to-mate), while Syzygybases use a variation of the DTZ metric (distance-to-zero, zero meaning any move that resets the 50-move counter). This special metric is one of the reasons that Syzygybases are more compact than Nalimov tablebases, while still storing all information needed for optimal play and in addition being able to take into account the 50-move rule.

Large Pages

Stockfish supports large pages on Linux and Windows. Large pages make the hash access more efficient, improving the engine speed, especially on large hash sizes. Typical increases are 5..10% in terms of nodes per second, but speed increases up to 30% have been measured. The support is automatic. Stockfish attempts to use large pages when available and will fall back to regular memory allocation when this is not the case.

Support on Linux

Large page support on Linux is obtained by the Linux kernel transparent huge pages functionality. Typically, transparent huge pages are already enabled and no configuration is needed.

Support on Windows

The use of large pages requires "Lock Pages in Memory" privilege. See Enable the Lock Pages in Memory Option (Windows) on how to enable this privilege, then run RAMMap to double-check that large pages are used. We suggest that you reboot your computer after you have enabled large pages, because long Windows sessions suffer from memory fragmentation which may prevent Stockfish from getting large pages: a fresh session is better in this regard.

Compiling Stockfish yourself from the sources

Stockfish has support for 32 or 64-bit CPUs, certain hardware instructions, big-endian machines such as Power PC, and other platforms.

On Unix-like systems, it should be easy to compile Stockfish directly from the source code with the included Makefile in the folder src. In general it is recommended to run make help to see a list of make targets with corresponding descriptions.

    cd src
    make help
    make net
    make build ARCH=x86-64-modern

When not using the Makefile to compile (for instance with Microsoft MSVC) you need to manually set/unset some switches in the compiler command line; see file types.h for a quick reference.

When reporting an issue or a bug, please tell us which version and compiler you used to create your executable. These informations can be found by typing the following commands in a console:

    ./stockfish compiler

Understanding the code base and participating in the project

Stockfish's improvement over the last couple of years has been a great community effort. There are a few ways to help contribute to its growth.

Donating hardware

Improving Stockfish requires a massive amount of testing. You can donate your hardware resources by installing the Fishtest Worker and view the current tests on Fishtest.

Improving the code

If you want to help improve the code, there are several valuable resources:

Terms of use

Stockfish is free, and distributed under the GNU General Public License version 3 (GPL v3). Essentially, this means that you are free to do almost exactly what you want with the program, including distributing it among your friends, making it available for download from your web site, selling it (either by itself or as part of some bigger software package), or using it as the starting point for a software project of your own.

The only real limitation is that whenever you distribute Stockfish in some way, you must always include the full source code, or a pointer to where the source code can be found. If you make any changes to the source code, these changes must also be made available under the GPL.

For full details, read the copy of the GPL v3 found in the file named Copying.txt.