Swiss Tournament Programs

[Krzysztof Grzelak]

In the near future, I would like to conduct a Swiss tournament among programs in international draughts. Any program can take part in the tournament. Information about the tournament - game time - 15 minutes for 75 moves + 5 seconds per move, rounds 9. The following programs will participate in the tournament:

Ares 1.91 - Joost Buijs - Netherlands
Cerberus 6.11 - Leo Nagels - Netherlands
Damage 16.1 - Bert Tuyt - Netherlands
Dragon Pro 4.6.2 - Michel Grimminck - Netherlands
Dream Gaïa 1.4 - Saïd Koudache - France
Edeon Sport 5.4.312 - Alexey Odnoklubov, Vasily Naumik, Victor Naumik - Belarus
Flits 3.02 - Adri Vermeulen - Netherlands
GWD 4.01 - Gijsbert Wiesenekker - Netherlands
Kingsrow 1.63 - Edgar Gilbert - United States of America
Maximus 2.02 - Jan-Jaap van Horssen - Netherlands
Moby Dam (Jul 12 2023) - Harm Jetten - Netherlands
Plus 500 6.56.2 - Serge Startsev - Russia
Scan 3.1 - Fabien Letouzey - France
Sjende Blyn 2026 - Jelle Wiersma - Netherlands
Tornado 7.26b - Frank Mesander - Netherlands
Truus voor Windows 1.0 - Stef Keetman - Netherlands

Unfortunately, I do not yet know when the tournament will begin as well as end. I will inform about everything on the forum. If any programmer would like to share the latest version of his program - you are welcome and thank you in advance. For more information about the tournament, visit https://www.swisstournament2026.draught ... eGames.php and https://results.fmjd.org/tournaments/2026/f_2260/. During the tournament I will use the program DraughtsArbiter Pro. The tournament is played on two identical computers with the following hardware parameters:

processor - AMD Ryzen Threadripper 2990WX 3.0 GHz
hard drive - Samsung 970 EVO Plus NVMe M.2 SSD 2 TB
memory ram - Kingston HyperX Fury 128 GB DDR4 2400 MHz CL15
monitor - Nec MultiSync EA245 WMi
keyboard - A4Tech KV-300H Slim
mouse - Razer DeathAdder Essential
operating system - Microsoft Windows 11 Professional 64 bit

Engine setting during the tournament:

Ares NNUE - Joost Buijs - Netherlands
Book - On
Threads - 32
TT-size - 32
Bucket-size - 4
Large memory pages - On
Pondering - On
The program uses from full 6 figure base of endings

Cerberus - Leo Nagels - Netherlands
Threads - 2
Book - On
Pondering - On
The program uses from full 6 figure base of endings

Damage NNUE - Bert Tuyt - Netherlands
Book - 0n
Threads - 32
NNUE-file - nn_20210328.gnn
TT-persistent - On
TT-size - 24
BB-cache - 5
BB-size - 7
BB-preload - 7
The program uses from full 7 figure base of endings

Dragon - Michel Grimminck - Netherlands
Max hashtable size - 16384 mb
Big evaluation table - 1243 mb
Opening book - Tournament book
Number of cores - 16
Permanent brain - On
Avoid draws - Not
The program uses from not full 8 figure base of endings

Dream - Saïd Koudache - France
Threads - 1
Utiliser bibliothčque d'ouvertures - On
Utiliser bibliothčque de finales - On
Réflexion sus temps adverse - On
Taille en Ko, pour les Hash Tables - 1024
The program does not use base of endings

Edeon Sport - Alexey Odnoklubov, Vasily Naumik, Victor Naumik - Belarus
Threads - 1
Use debuts book - On
The program uses from not full 8 figure base of endings

Flits - Adri Vermeulen - Netherlands
Threads - 1
Transpositietabel(Mb) - 1770
Pondering - On
The program does not use base of endings

GWD - Gijsbert Wiesenekker - Netherlands
Threads - 32
Hashtable size - 16384
Opening book - On
Pondering - On
The program uses from full 7 figure base of endings

Kingsrow - Edgar Gilbert - United States of America
Hashtable size (mb) - 512
Cache size (mb) - 30000
Max.pieces - 8
Search threads - 16
Opening book - Best moves
Pondering - On
The program uses from full 8 figure base of endings

Maximus - Jan-Jaap van Horssen - Netherlands
Threads - 8
Opening book - On
Book margin - 0.01
Pondering - On
Transposition Table - 32 M positions (25 bits)
The program uses from full 6 figure base of endings

Moby Dam - Harm Jetten - Netherlands
Opening book - Off
Threads - 8
TT-size - 28
Pondering - On
Large memory pages - On
The program uses from full 6 figure base of endings

Plus 500 - Serge Startsev - Russia
Threads - 1
The openings book - On
Self-learning - On
Endgames DB - On
The program uses from full 6 figure base of endings

Scan - Fabien Letouzey - France
Hub - On
Variant - Normal
Book - On
Book ply - 4
Book-margin - 4
Pondering - On
Threads - 16
TT-size - 26
BB-size - 6
The program uses from full 6 figure base of endings

Sjende Blyn 2026 - Jelle Wiersma - Netherlands
Threads - 32
Strength - Maximum
Pondering - On
The program uses from full 6 figure base of endings

Tornado - Frank Mesander - Netherlands
Threads - 5
Pondering - On
Hash Table Geheugen - 2G
The program uses from full 6 figure base of endings

Truus voor Windows - Stef Keetman - Netherlands
Threads - 1
Pondering - On
Partijen database ruimte (MB) - 64
6 Stukken Database (anders 5) - On
Stellingen ruimte (16K) - 10000
The program uses from full 6 figure base of endings

If anyone has any questions - please ask - I will answer.

Krzysztof

[Krzysztof Grzelak]

Sorry to ask if any of the programmers have a new version of the program playing in the tournament.

Krzysztof

[FrankMesander]

Sorry to ask if any of the programmers have a new version of the program playing in the tournament.

Krzysztof

I am going to send you a new version of Tornado, but because I am still working on it, I will send you the most recent version when I know when the tournament will start.

Frank.

[pontel]

Before making my constructive comment, I would first like to praise and thank Grzelak for the work he has done so far with his tournaments. I am an enthusiast of draughts engines as well as of his tournaments, and I even make use of this work by recording the games that did not end in draws in my database.

I understand that only tournament formats similar to TCEC (Top Chess Engine Championship) can truly reveal the best engine. Round-robin or Swiss formats with only one or two games per pairing usually end up declaring the champion mainly based on performance against the weaker programs, since the strongest engines often draw against each other.

One interesting suggestion would be to start with a Swiss or round-robin system, then have the top four programs qualify for the semifinals, followed by a final consisting of a multi-game match between the finalists. Of course, such an undertaking would depend on having an efficient DXP system.

[Ed Gilbert]

In American 8x8 checkers, when they realized there were too many draws, they started using more difficult start positions. First they used 2-move openings, and then later changed to 3-move openings. For 10x10 draughts, a similar idea might be to use a set of difficult start positions to get more decisive games between the stronger engines.

-- Ed