New Chess Tournament

[Krzysztof Grzelak]

Intel Core I7 2670QM
System - Windows Home Premium 64 Bit (SP1)
GUI - ChessGUI 0.245g
Hash - 1024 MB
TB Cache - 128 MB
Each Hash Engine - 1024 MB
Book - Baracuda2014c.cgb - 8 moves
Tablebases - 5 Namilov, 5 Gaviota
Time - 30 minutes per game
Ponder - OFF
All engine CPU - 1
Tournament type - Swiss
Number of rounds - 11

Participants

Alfi 13.1 x64
Arminus 2014-01-18 x64
Capivara LK 0.09b02d x64
Coiled 0.02b x64
Delphil 3.2 x64
Hamsters 0.7.1 x64
Ice 2.0 v2240 x64
LEOpard 0.7c x64
MinkoChess 1.3 x64
Nebula 2.0 x64
Octochess r.5190 x64
Predateur 2.2.1 x64
RedQueen 1.1.97 x64
Rhetoric 1.4.1 x64
Toga II 28.05.2013 x64
Vajolet 2 1.45 x64

[Graham Banks]

A 5 round Swiss would be enough for such a small field.

As you're using 16 engines, a knockout might be worth thinking about.

[Krzysztof Grzelak]

Round 1 Pairings

Alfi 13.1 x64 - MinkoChess 1.3 x64
Nebula 2.0 x64 - Arminus 2014-01-18 x64
Capivara LK 0.09b02d x64 - Octochess r.5190 x64
Predateur 2.2.1 x64 - Coiled 0.02b x64
Delphil 3.2 x64 - RedQueen 1.1.97 x64
Rhetoric 1.4.1 x64 - Hamsters 0.7.1 x64
Ice 2.0 v2240 x64 - Toga II 28.05.2013 x64
Vajolet 2 1.45 x64 - LEOpard 0.7c x64

[Krzysztof Grzelak]

It thanks for valuable advices.

[Guenther]

Intel Core I7 2670QM
System - Windows Home Premium 64 Bit (SP1)
GUI - ChessGUI 0.245g
Hash - 1024 MB
TB Cache - 128 MB
Each Hash Engine - 1024 MB
Book - Baracuda2014c.cgb - 8 moves
Tablebases - 5 Namilov, 5 Gaviota
Time - 30 minutes per game
Ponder - OFF
All engine CPU - 1
Tournament type - Swiss
Number of rounds - 11

Participants

LEOpard 0.7c x64

Another shameless cloner who even maintained(in 2011/07) his new chess program is 'independent' and not derived from any others...
The author is an anonymous called after R. Daneel Olivaw who is a fictional robot created by Isaac Asimov.

The investigation below is found via google:

e.g.
http://immortalchess.net/forum/archive/ ... 10972.html

http://www.chess2u.com/t3161-really-str ... s-07c-beta

(BTW it seems it is exactly Ivanhoe B47)

Code: Select all

the LEOPard website
http://www.giocareascacchi.it/utility/forum/Chess-Engine/1986-LEOpard-U-C-I-chess-engine-.html/1/

translated via google:
Leopard is an engine "independent" in that it is not derived from any existing engine, even though I took many ideas and procedures from several open source engines, the first Ivanhoe stockfish and foremost.

strings from IvanHoe that can be found in the LEOpard executable:

NEW_PONDERHIT
INCREASED_DEPTH
time_factor
EASY
BATTLE
ORDINARY
The GUI with you owns not strict FEN compliantcy. Demand anew!
FEN col: %d tra: %d pez: %d
Button %s pushed, action worked %s

there's more but this is enough for me.
seems clearly based on IvanHoe...

Perhaps he should have said:
"Without many ideas and techniques from the open source chess engines Ippolit and Stockfish, LEOpard would not nearly be as strong as it is now."

too...

SMP_HAS_AKTIV
UCI_White_Bishops_Scale
UCI_White_Pawn_Scale
UCI_White_Knight_Scale
UCI_White_Light_Scale
UCI_White_Dark_Scale
UCI_White_Rook_Scale
UCI_White_Queen_Scale
UCI_Black_Bishops_Scale
UCI_Black_Pawn_Scale
UCI_Black_Knight_Scale
UCI_Black_Light_Scale
UCI_Black_Dark_Scale
UCI_Black_Rook_Scale
UCI_Black_Queen_Scale
UCI_MATERIAL_WEIGHTING
UCI_PAWNS_WEIGHTING
UCI_KING_SAFETY_WEIGHTING
UCI_STATIC_WEIGHTING
UCI_MOBILITY_WEIGHTING

IVAN_INIT
IVAN_SPLIT
ivan_thread
ivan_init_smp
ivan_fail_high
ivan_end_smp
SMP_IVAN
IVAN_ALL_HALT

and more...

Another note:
It is said it uses all available CPUs if SMP is not disabled explicitely(thread option is removed to appear stronger in tests).
This means it will cheat in your test if you don't follow step 1.

[Krzysztof Grzelak]

Another note:
It is said it uses all available CPUs if SMP is not disabled explicitely(thread option is removed to appear stronger in tests).
This means it will cheat in your test if you don't follow step 1.

I'm sorry that I ask, what do you suggest in this case.These are things I did not know about this engine.

[Guenther]

Another note:
It is said it uses all available CPUs if SMP is not disabled explicitely(thread option is removed to appear stronger in tests).
This means it will cheat in your test if you don't follow step 1.

I'm sorry that I ask, what do you suggest in this case.

Like I said above you must disable SMP in the uci options otherwise it will use all cpus as it does not use the standard threads option.
Then check with your taskmanager if it really works with one cpu according to your tournament conditions.
This should be done anyway for all participants to avoid errors in the setup. It would be disastrous to have one playing on 4 cpus instead of 1.

(BTW if you advertise tournaments here it would be also good to include evals/move time and prepare a download of the games)

Much better though would be to replace it by another program.
There is not much sense in using a 3 years old known clone and publish results of it in CCC.
OTH it is your tournament and you must decide what your goal is.


Guenther

[Krzysztof Grzelak]

Like I said above you must disable SMP in the uci options otherwise it will use all cpus as it does not use the standard threads option.
Then check with your taskmanager if it really works with one cpu according to your tournament conditions.
This should be done anyway for all participants to avoid errors in the setup. It would be disastrous to have one playing on 4 cpus instead of 1.

(BTW if you advertise tournaments here it would be also good to include evals/move time and prepare a download of the games)

Much better though would be to replace it by another program.
There is not much sense in using a 3 years old known clone and publish results of it in CCC.
OTH it is your tournament and you must decide what your goal is.


Guenther

Thank you for the advice. Barter it on another engine and the tournament will begin from new.

Krzysztof.

[Matthias Gemuh]

A 5 round Swiss would be enough for such a small field.

As you're using 16 engines, a knockout might be worth thinking about.

Krzysztof, the formula is always "square root of number of participants, rounded _up_ to next odd number".

In ChessGUI, I recommend Pairing Type 2 for Swiss (algorithm of Volker Annuss).

Matthias.

[Krzysztof Grzelak]

Krzysztof, the formula is always "square root of number of participants, rounded _up_ to next odd number".

In ChessGUI, I recommend Pairing Type 2 for Swiss (algorithm of Volker Annuss).

Matthias.

Thank you so I have checked in the program. I have an important question about the program ChessGUI. Please write what to do if one of the engines oneself hangs at the tournament. Do you need to re-play tournament.