Battle of the Forgotten

Discussion of computer chess matches and engine tournaments.

Moderators: hgm, Rebel, chrisw

User avatar
hgm
Posts: 27790
Joined: Fri Mar 10, 2006 10:06 am
Location: Amsterdam
Full name: H G Muller

Re: Battle of the Forgotten

Post by hgm »

Carlos777 wrote:Not working under Arena: (if someone has an idea on how to make them work, please reply)

ChessV 0.94
The engine command line for ChessV 0.94 has to be:

Code: Select all

ChessV_WinBoard "Orthodox Chess"
It seems that Arena cannot handle this (or people simply have not found out what they have to tell Arena to produce this command line when installing the engine).

Possible work-arounds are:

1) Put that command line in a batch file chessv.bat, and then install that batch file as an engine in Arena.

2) Make a ChessV .SGF file (saved-game file), say CHESS.SGF for the initial position of Orthodox Chess by saving from the stand-alone version of ChessV, or read the instructions that come with ChessV on how to make such a file for specifying variants, and follow those. Then install ChessV in Arena with the engine command:

Code: Select all

ChessV_WinBoard CHESS.SGF
so that you do not need the quotes around the argument that specifies the variant.

Even under WinBoard + PSWBTM I have to use the latter trick to make ChessV play Bird's Chess, as I have found no way yet to get the single quote in the variant name into the engine command line. With a bare WinBoard I can use /fcp={ChessV_WinBoard "Bird's Chess"}, but PSWBTM puts single quotes around the /fcp argument, and that is what makes WinBoard choke on it.
metax
Posts: 344
Joined: Wed Sep 23, 2009 5:56 pm
Location: Germany

Re: Battle of the Forgotten

Post by metax »

What is the approximate rating scale for '1700 or lower'?

In CCRL rating, I would estimate ChessMind at 1930-1950 Elo based on games against Surprise. But in other rating scales, the ratings are often lower.

However, based on games against humans, I would estimate ChessMind at 1700-1750 'human Elo'.

I think in the CCRL rating scale Jabba would have around 1980-2000 Elo or even more.
Adam Hair
Posts: 3226
Joined: Wed May 06, 2009 10:31 pm
Location: Fuquay-Varina, North Carolina

Re: Battle of the Forgotten

Post by Adam Hair »

This is great Carlos. I've been working on a rating list since August
and I am including engines of all strengths. Here is the bottom of my
list. I plan to include as many engines as I can make run on my computer
under Arena 2.01.

Code: Select all

77 Warrior_103                   -214   28   28   984   54%  -246   20% 
  78 Smash                         -248   27   27  1063   49%  -237   16% 
  79 SnailChess_4013               -252   29   29  1037   66%  -398   15% 
  80 Surprise_4.3b13               -254   28   28   970   49%  -244   16% 
  81 Clueless_14                   -278   28   28   984   46%  -241   11% 
  82 Embracer_112                  -309   28   28  1023   60%  -394   28% 
  83 Grizzly1401b                  -334   29   29  1023   57%  -395   13% 
  84 Elf130                        -403   29   29  1065   50%  -401   16% 
  85 TSCP181                       -430   28   28  1154   51%  -438   15% 
  86 Mizar3.0                      -453   30   30  1015   53%  -477   18% 
  87 DChess1_0_2                   -469   30   30  1018   51%  -475   16% 
  88 Storm_06                      -485   29   29  1034   49%  -473   17% 
  89 Beaches226                    -494   31   31   901   52%  -507   14% 
  90 PolarChess1.2                 -500   30   30   986   47%  -475   17% 
  91 Bace_045                      -513   29   29  1050   46%  -473   17% 
  92 MSCP_144_uci_ja               -517   31   31   981   51%  -553   18% 
  93 Needle_0531                   -563   31   31   872   53%  -618   23% 
  94 LarsenVB                      -638   32   32   877   45%  -616   15% 
  95 Pentagon_12                   -657   31   31   869   35%  -539   28% 
  96 StAndersen_v1.31_X64          -675   32   32   867   34%  -539   15% 
  97 TarraschToyEngineV0.906       -681   31   31   942   32%  -531   15% 
  98 Yawce016                      -687   34   34   770   45%  -660   10% 
  99 Nero_61                       -717   32   32   961   33%  -571   16% 
 100 Numpty_0.4pr                  -924   40   40   570   35%  -807   17% 
 101 Marquis16                     -951   45   45   536   50%  -953   16% 
 102 Xadreco_v5.7                 -1192   57   57   420   31%  -956   13% 
 103 Testina2_2                   -1259   64   64   241   34% -1096   31% 
 104 Pyotr-am_06                  -1376   63   63   387   12%  -947   18% 
The only engine in this group that gives me problems is Testina2.2. It
loses some on time and sometimes makes illegal moves.
Carlos777
Posts: 1730
Joined: Sun Dec 13, 2009 6:09 pm

Re: Battle of the Forgotten

Post by Carlos777 »

hgm wrote:
Carlos777 wrote:Not working under Arena: (if someone has an idea on how to make them work, please reply)

ChessV 0.94
The engine command line for ChessV 0.94 has to be:

Code: Select all

ChessV_WinBoard "Orthodox Chess"
Hi HGM,

I tried adding "Orthodox Chess" in the command parameters box without success, I will try your suggestions when I back home.

Thanks,
Carlos
User avatar
hgm
Posts: 27790
Joined: Fri Mar 10, 2006 10:06 am
Location: Amsterdam
Full name: H G Muller

Re: Battle of the Forgotten

Post by hgm »

Yes, people told me that Arena cannot handle quotes in the engine command line at all. But the sgf trick should work. I just tried it to play "Chess With Different Armies", because you could not even specify that from the command line without a new popup appearing to ask you which armies you want to pit against each other.

So what I did is just start a game of Chess with Different Armies by hand, select FIDE vs Color-bound Clobberers, and then immediately select "File -> Save Game". By giving the name of the save file as an arument, ChessV then plays that variant without problems under WinBoard + PSWBTM.
Carlos777
Posts: 1730
Joined: Sun Dec 13, 2009 6:09 pm

Re: Battle of the Forgotten

Post by Carlos777 »

metax wrote:What is the approximate rating scale for '1700 or lower'?
I know that it is kind of ambiguous. As few tests have been done with low ELO engines (UEL is one of the few testing low ELO UCI engines, Open War and WBEC have done it too), there is no certainty about what engines should be considered 1700 ELO or less. Anyway, when I thought about 1700 ELO, I was thinking about human ELO.
However, based on games against humans, I would estimate ChessMind at 1700-1750 'human Elo'.
In that case, I'd be glad to test ChessMind in the next cycle if you agree.
I think in the CCRL rating scale Jabba would have around 1980-2000 Elo or even more.
Yes, as I said before Jabba is probably too strong for this tournament. I did not know its strenght before testing.
Last edited by Carlos777 on Tue Jan 19, 2010 4:49 am, edited 1 time in total.
Carlos777
Posts: 1730
Joined: Sun Dec 13, 2009 6:09 pm

Re: Battle of the Forgotten

Post by Carlos777 »

Adam Hair wrote:This is great Carlos. I've been working on a rating list since August
and I am including engines of all strengths. Here is the bottom of my
list. I plan to include as many engines as I can make run on my computer
under Arena 2.01.

Code: Select all

77 Warrior_103                   -214   28   28   984   54%  -246   20% 
  78 Smash                         -248   27   27  1063   49%  -237   16% 
  79 SnailChess_4013               -252   29   29  1037   66%  -398   15% 
  80 Surprise_4.3b13               -254   28   28   970   49%  -244   16% 
  81 Clueless_14                   -278   28   28   984   46%  -241   11% 
  82 Embracer_112                  -309   28   28  1023   60%  -394   28% 
  83 Grizzly1401b                  -334   29   29  1023   57%  -395   13% 
  84 Elf130                        -403   29   29  1065   50%  -401   16% 
  85 TSCP181                       -430   28   28  1154   51%  -438   15% 
  86 Mizar3.0                      -453   30   30  1015   53%  -477   18% 
  87 DChess1_0_2                   -469   30   30  1018   51%  -475   16% 
  88 Storm_06                      -485   29   29  1034   49%  -473   17% 
  89 Beaches226                    -494   31   31   901   52%  -507   14% 
  90 PolarChess1.2                 -500   30   30   986   47%  -475   17% 
  91 Bace_045                      -513   29   29  1050   46%  -473   17% 
  92 MSCP_144_uci_ja               -517   31   31   981   51%  -553   18% 
  93 Needle_0531                   -563   31   31   872   53%  -618   23% 
  94 LarsenVB                      -638   32   32   877   45%  -616   15% 
  95 Pentagon_12                   -657   31   31   869   35%  -539   28% 
  96 StAndersen_v1.31_X64          -675   32   32   867   34%  -539   15% 
  97 TarraschToyEngineV0.906       -681   31   31   942   32%  -531   15% 
  98 Yawce016                      -687   34   34   770   45%  -660   10% 
  99 Nero_61                       -717   32   32   961   33%  -571   16% 
 100 Numpty_0.4pr                  -924   40   40   570   35%  -807   17% 
 101 Marquis16                     -951   45   45   536   50%  -953   16% 
 102 Xadreco_v5.7                 -1192   57   57   420   31%  -956   13% 
 103 Testina2_2                   -1259   64   64   241   34% -1096   31% 
 104 Pyotr-am_06                  -1376   63   63   387   12%  -947   18% 
The only engine in this group that gives me problems is Testina2.2. It
loses some on time and sometimes makes illegal moves.
Thanks Adam for this useful information, many of those engines will be tested in the next cycles.

OTOH, I am considering to increase the number of engines to be removed after the cycle ends to 20 (instead of 15) to speed things up. More engines would be tested in this way (new 20 engines.) If someone disagree, let me know your reason.

Regards,
Carlos
JVMerlino
Posts: 1357
Joined: Wed Mar 08, 2006 10:15 pm
Location: San Francisco, California

Re: Battle of the Forgotten

Post by JVMerlino »

Hi Carlos,

Thanks for including Myrddin in your tournament. The version you have chosen was rated 1600-1700, so it does fit your tournament's criteria. However, Myrddin is definitely still in development, and the latest version is about 250-300 points stronger than the version in your tournament.

Someday, Myrddin will indeed be forgotten. :D But hopefully not for a while yet. :wink:

Thanks again!
jm
Carlos777
Posts: 1730
Joined: Sun Dec 13, 2009 6:09 pm

Re: Battle of the Forgotten

Post by Carlos777 »

JVMerlino wrote:Hi Carlos,

Thanks for including Myrddin in your tournament. The version you have chosen was rated 1600-1700, so it does fit your tournament's criteria. However, Myrddin is definitely still in development, and the latest version is about 250-300 points stronger than the version in your tournament.

Someday, Myrddin will indeed be forgotten. :D But hopefully not for a while yet. :wink:

Thanks again!
jm
Hi John,

You are welcome. As I wrote in the first post, Myrddin was one of the first engines that I started testing to find a suitable opponent for me. I used this old version on purpose, sorry. :) When I decided to run a tournament for testing more low ELO engines, I chose to keep it given that Myrddin 0.081a2 had already played many games.

Myrddin 0.82 would be too strong for this tournament and it looks like this "weaker" version is going to be in the top 10 at the end of the current cycle.

I don't think that Myrddin will be forgotten. :wink:

Regards,
Carlos
JVMerlino
Posts: 1357
Joined: Wed Mar 08, 2006 10:15 pm
Location: San Francisco, California

Re: Battle of the Forgotten

Post by JVMerlino »

Carlos777 wrote:Hi John,

You are welcome. As I wrote in the first post, Myrddin was one of the first engines that I started testing to find a suitable opponent for me. I used this old version on purpose, sorry. :) When I decided to run a tournament for testing more low ELO engines, I chose to keep it given that Myrddin 0.081a2 had already played many games.

Myrddin 0.82 would be too strong for this tournament and it looks like this "weaker" version is going to be in the top 10 at the end of the current cycle.

I don't think that Myrddin will be forgotten. :wink:

Regards,
Carlos
Thanks for the kind words. Yes, the most recent public release is about 300 ELO stronger than 0.81, so way too strong for this tournament. I've had a few people say that they enjoy playing it OTB, which is very nice to hear. It's also very stable (so far) according to various tournament directors like yourself.

Thanks again!

jm