The Baron 3.44 (New Version)

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

User avatar
Steve Maughan
Posts: 1221
Joined: Wed Mar 08, 2006 8:28 pm
Location: Florida, USA

The Baron 3.44 (New Version)

Post by Steve Maughan »

Richard Pijl has launched a new version of The Baron. This fixes some Syzygy bug and generally tunes the engine. Richard estimate the improvement in strength is between 20 ELO and 50 ELO.

Find out more and download here:

http://www.chessprogramming.net/the-baron/

Steve
http://www.chessprogramming.net - Maverick Chess Engine
Modern Times
Posts: 3546
Joined: Thu Jun 07, 2012 11:02 pm

Re: The Baron 3.44 (New Version)

Post by Modern Times »

I'll have to try it with chess960. That last version that worked for me was 3.29; I could not get any version after that to play it successfully.
User avatar
Guenther
Posts: 4605
Joined: Wed Oct 01, 2008 6:33 am
Location: Regensburg, Germany
Full name: Guenther Simon

Re: The Baron 3.44 (New Version)

Post by Guenther »

Steve Maughan wrote: Sun Dec 30, 2018 9:38 pm Richard Pijl has launched a new version of The Baron. This fixes some Syzygy bug and generally tunes the engine. Richard estimate the improvement in strength is between 20 ELO and 50 ELO.

Find out more and download here:

http://www.chessprogramming.net/the-baron/

Steve
Thanks to Richard for the new Baron and a happy new year for you both.
https://rwbc-chess.de

trollwatch:
Chessqueen + chessica + AlexChess + Eduard + Sylwy
Damir
Posts: 2801
Joined: Mon Feb 11, 2008 3:53 pm
Location: Denmark
Full name: Damir Desevac

Re: The Baron 3.44 (New Version)

Post by Damir »

Thanks for the new Baron Richard. Happy New Year to both of you. :) :) :D
Modern Times
Posts: 3546
Joined: Thu Jun 07, 2012 11:02 pm

Re: The Baron 3.44 (New Version)

Post by Modern Times »

I might be having some success with FRC with the KQkq_OO castling setting under newer versions of ChessGUI, as a UCI engine. It refuses to play under Shredder GUI and under older ChessGUI versions. I'll report back further. If anyones cares :mrgreen:
Modern Times
Posts: 3546
Joined: Thu Jun 07, 2012 11:02 pm

Re: The Baron 3.44 (New Version)

Post by Modern Times »

Unfortunately I give up on FRC support with this, as I've had to do on the previous 3.4x versions. No setting seems to work with it.
pijl
Posts: 115
Joined: Mon Sep 17, 2012 8:59 pm

Re: The Baron 3.44 (New Version)

Post by pijl »

I'll take a look at it. I'm not aware of making changes that could break chess960 support, but apparently something is not working as expected. I must confess I haven't used it myself in a while.
Richard.
Modern Times
Posts: 3546
Joined: Thu Jun 07, 2012 11:02 pm

Re: The Baron 3.44 (New Version)

Post by Modern Times »

The only thing I can see in the logs is this:

WARNING: Castling for white specified, but no rook at the H-file.

I use 3.29 as a winboard engine, but I don't know whether that or UCI is the best choice for 3.44.
User avatar
Werner
Posts: 2871
Joined: Wed Mar 08, 2006 10:09 pm
Location: Germany
Full name: Werner Schüle

Re: The Baron 3.44 (New Version)

Post by Werner »

Hi, I use this .ini file (5men s bases on SSD) or Options set in uci mode:
- what About Kingsavety eval or razoring ??

# Configuration file of the Baron

# Amount of Transposition Hashtables in MB
mainhash 256

# Learning, on by default
# Allowed values: on/off
# Recommended value: on
learning off

# Name and location of the opening book
# A total of 16 books in baron and polyglot format can be used. The books will be probed
# in the order of their listing here.
# bookfile: Book in Baron format
# bookpoly: Book in Polyglot format
# bookclient: Accessing remote boot on Baron's book server (will require additional license in the future)
# bookctg: Book in Chessbase format (note, support is not complete, use with care)
#bookclient 127.0.0.1:4463
#bookfile d:/baron/books/big
#bookpoly d:/baron/books/general.bin
#bookctg d:/baron/books/baron.ctg

# Whether nullmove is used or not
# Allowed values: on/off
# Default/Recommended: on
# nullbase: number of quarterplies used for nullmove reduction
# nullprogress: measure of progressive nullmove reduction based on depth (more reduction for deeper searches)
donullmove on
nullbase 12
nullprogress 6

# Internal Iterative Deepening.
# Allowed values: on/off
# Default/Recommended: on
doiid on

# Singular extensions
# Extension parameters specified in quarter plies
# Default/Recommended values:
# singular on
# singreduction 4
# singmargin 100
singular on
singmargin 100
singextension 4

# LMR
# Specified in quarter plies
# Default/Recommended values:
# lmrpv 0
# lmrscout 4
# lmrextrapv 6
# lmrextrascout 2
#lmrpv 0
#lmrscout 4
#lmrextrapv 4
#lmrextrascout 4

# Razoring
# Specified in full plies
#prunelevel 4
#prunepv off
#prunesecure 100

# The next settings are only meant for test/tuning purposes. For 'regular' use the parameter 'prunesecure' should be used instead.
# prunealphamargin <depthleft> <percentage of default> specifies the amount the quiescent value should be below alpha so that the
# quiescent score+margin will be returned by the search instead of searching it to depthleft
# prunebetamargin <depthleft> <percentage of default> specifies the amount the quiescent value should be above beta so that the
# quiescent score-margin will be returned by the search instead of searching it to depthleft
#prunealphamargin 1 100
#prunebetamargin 1 100
#prunealphamargin 2 100
#prunebetamargin 2 100
#prunealphamargin 3 100
#prunebetamargin 3 100
#prunealphamargin 4 100
#prunebetamargin 4 100
#prunealphamargin 5 100
#prunebetamargin 5 100
#prunealphamargin 6 100
#prunebetamargin 6 100
#prunealphamargin 7 100
#prunebetamargin 7 100
#prunealphamargin 8 100
#prunebetamargin 8 100

# King Safety Evaluation
# kingsafety scores the attack against the king (offensive)
# kingshield scores the shield in front of the king (defensive)
# Percentage of the default scoring method
#kingsafety 100
#kingshield 100

# Multi processor configuration:
# cpus: Number of CPU's used (max 16 or the number of processors in your system, whatever is the lower number)
# splitdepth: number of quarter-plies that needs to be remaining to attempt to split the search.
# minsplit: number of remaining moves in a position to attempt creating a splitpoint (at least 2)
# maxsplit: maximum number of threads that may be assigned to a splitpoint (default: all)
# gluecores: 'Glue' the workers to a specific cpu core (only set to 'on' when the Baron has exclusive use of the machine)
cpus 1
splitdepth 12
minsplit 2
maxsplit 32
gluecores off

# Keep the noise level down
#postoutput 1

# Some more options
# timeuse: Percentage of default amount of time allocated to make a move
# easymove: Reply faster when there is one obvious good move
# logging: Turns on/off the logfile (has the name of this configuration file with the extension .log)
# priolow: Sets the engine to run at lower priority to improve responsiveness of the system
# ucioption (UCI): Allows setting options through UCI option commands. Setting it to off will still advertise the options,
# but the engine will ignore the commands. Values from this file will be used for all.
# icsmode (WB): allows the engine to select between icsmode (which will kibitz/whisper information). Values are 'on', 'off' and 'smart'
# When set to 'smart' the engine will determine ics mode based on the winboard ics command.
#timeuse 100
#easymove on
logging off
priolow off
ucioption on
#icsmode smart

# Syzygy Base options
# syzygypath: path to where the sygyzybases are located.
# szqsearch: Turns probing on/off in quiescense search. Only turn on with bases on SSD or flashdisk
# szdepth: Minimum remaining depth in quarterplies to probe the bases. Must be 0 to probe in qsearch
syzygypath c:\sbases
szqsearch on
szdepth 0
Werner
pijl
Posts: 115
Joined: Mon Sep 17, 2012 8:59 pm

Re: The Baron 3.44 (New Version)

Post by pijl »

Hi Werner,
Not sure what your question is. The commented out values are the defaults, but you can experiment with them. If you find other values to be clearly stronger I'd appreciate the feedback.
I see you enabled Syzygy in qsearch, which can be really heavy, especially on multicore machines, but for single core it may be ok.
Richard.