WyldChess Latest

This is a release of the current master branch.

While this release contains features such as Lazy SMP, FRC and some strength improvements, I will not request the CCRL team to test this specifically as a new version of WyldChess unless they wish to do so. The code is 3+ years old and this repository is no longer being actively developed. I am, however, open to fixing bugs and such.

Instructions for using the Persona feature

Use the UCI option: PersonaPath and provide the full path to a persona file (for example drunk_attacker.txt provided in the release here).
The persona file is a preset of UCI evaluation parameters, so it is quite simple to create one.

For example:
setoption name PersonaPath value C:\Users\user\drunk_attacker.txt

WyldChess 1.51

This is a patch release for WyldChess 1.5.

It contains the patches for the following bugs:

• The Xboard mode bug where it would play illegal moves when starting from a given FEN/EPD.
• Windows timeouts (by increasing default move overhead to 30ms instead of previous 10ms).

It also contains a new UCI/Xboard option: MoveOverhead, which allows the user to set the amount of time WyldChess should reserve from it's usual allotment of time for each move. This allows further customization in case the 30ms reservation proves too little.

For ChessGUI users, if increasing the MoveOverhead still causes time forfeits, a suggestion from Norbert Raimund Leisner:

To avoid problems with time forfeits, you can change the default setting: ChessGUI->Game->Game Options->GUI1->TC overstep margin = 5000ms

WyldChess 1.5

Changes:
-> Added Syzygy EGTB support
-> Improved LMR (in part thanks to Ed Schroeder!)
-> Added history heuristic
-> Fixed TT mate-score bugs
-> Better futility pruning
-> Begun switching to c++ for easier coding(and possibly smaller binaries for windows)
This version I mainly modified search without changing evaluation much. Self testing shows significant(100+) elo gains at 40moves/90s and a little less at 40/10s+0.1s but I doubt it will work out better than ~50elo going by the CCRL list.

Next version I intend to switch completely to C++, actually improve the evaluation and finally add multi-threading support.

EDIT June 18, 2017: Updated Android release to fix the infinite thinking bug.

WyldChess 1.4

Changes since 1.3:
-> Fixed time control bugs where the increment was greater than the total time left
-> Reversed futility pruning(now in the move loop)
-> Fixed check extension bug
-> Improved(and simplified) passed pawn evaluation
-> Refactored move ordering
-> Reverted to old LMR formula(log-based formula missing crucial moves occasionally)
Should be ~50 elo stronger and easier to tinker with

Edits:
03/13/2017 - Fixed UCI crash bug

WyldChess 1.3

Needs fixing

WyldChess 1.2

Changes:
- Deterministic
- Fixed time per move setting
- Other miscellaneous bug fixes

WyldChess 1.1

Changes:
-Added analysis support
-Fixed xboard Illegal move/ERROR bug
-Supporting both '' and 'usermove ' in xboard mode now
-Finally have a decent makefile
-Added a file used to automatically build the releases

WyldChess 1.0

First public release