Winboard Forum

Hi !

As some or even most of you may know, the source of GnuChess 3.1+ is (and has always been ?) available at this site :
http://www.tim-mann.org/gnuchess.html

It is, if I am not mistaken, the first xboard/winboard programm ever !

Wouldn't it be nice to see how this 1991 champ fares against later competition ?

Could some kind soul compile it for modern systems, like (shameless self-interest) Windows XP ?

Many thanks in advance.

YS

I don't believe version 3.1 had Winboard support. That was first introduced in about 1993. The sources for 3.1 are for a Windows GUI program (and it looks like maybe it is Windows 3.1 or that vintage - doesn't compile on a modern system).

I tried the more commonly available Gnuchess 4.0 a while back and found it was quite weak against recent engines, even amateur ones, especially in endgame.

jdart wrote:I don't believe version 3.1 had Winboard support. That was first introduced in about 1993. The sources for 3.1 are for a Windows GUI program (and it looks like maybe it is Windows 3.1 or that vintage - doesn't compile on a modern system).

I tried the more commonly available Gnuchess 4.0 a while back and found it was quite weak against recent engines, even amateur ones, especially in endgame.

I've just added basic winboard support.
The 3.1 sources I have are for console based engine and it compiles ok with cygwin-gcc.
Once I've sorted out the time control I'll post a link to it.

Jim.

jdart wrote:I don't believe version 3.1 had Winboard support. That was first introduced in about 1993. The sources for 3.1 are for a Windows GUI program (and it looks like maybe it is Windows 3.1 or that vintage - doesn't compile on a modern system).
.

Hi

Guenther Simon's history pages say 1991 for the first release of GnuChess and in the readme of the source I've downloaded (link in the first post of this thread) one can read :

This is an update of gnuchess version 3.1.

It fixes all known bugs with pawn promotion, undo, and win/draw determination.

It adds the capability of reading xboard position files (xget).

It adds specification of save and load files on the command line (gnuchessr).

Save file format is now human readable.

'white' and 'black' no longer start play but only set the players. If the
computer is to move first use the 'go' command.


New commands are xget and hashlevel.

xget reads an xboard position file.

hashlevel allows you to specify parameters for using the hashfile.

So I believed it could already be winboard.

YS

jdart wrote:I tried the more commonly available Gnuchess 4.0 a while back and found it was quite weak against recent engines, even amateur ones, especially in endgame.

GnuChess 4.077 (1996) was quite strong for its time, but yes its endgame play is horrible (as for many other old ones). What more it has a bug in tournament control time and another one with uncorrect drawclaims.

Under the kinder (at least for it) rules of my dinosaur league (fisher clock, nunn positions, etc.) it stands quite well for its age, I think. Here's a little part of the standings :

Code: Select all

   2 Ruffian 1.01          2621   22   22   768   67%  2494   24%
   9 Crafty 17.14          2524   16   16  1536   64%  2420   25%
  30 SOS 99.11             2454   13   13  2240   58%  2394   20%
  43 Crafty 15.20          2418   13   13  2176   60%  2341   24%
  63 Phalanx 21            2340   17   17  1344   51%  2331   16%
  80 Arasan 5.3            2258   17   17  1280   41%  2324   21%
  81 ZChess 1.0            2257   15   14  1920   49%  2260   17%
  82 Bringer 1.5           2254   15   14  1856   52%  2239   19%
  88 GnuChess 4.077        2240   13   13  2240   47%  2261   20%
  99 InmiChess 3.01        2174   18   18  1152   47%  2193   17%
 106 Lambchop 6.90         2140   16   16  1536   40%  2220   15%

Best wishes.
YS

Jim Ablett wrote:I've just added basic winboard support.
The 3.1 sources I have are for console based engine and it compiles ok with cygwin-gcc.
Once I've sorted out the time control I'll post a link to it.

Jim.

Hi Jim,

Many thanks !
You are truly amazing !!

Best wishes.
YS

GnuChess 3.1 by J.Stanback/S.Cracraft (1991)

Winboard protocol added.
Supports time control > x moves in y mins.

Download:
http://www.mediafire.com/?4gz0vmzmmzc

Todo:
When I get time I will enable/correct the p.v output.



Jim.

Here's a new build of the original/unhacked by me, GnuChess 3.1 (console mode only)
version, which will run modern Windows systems. I thought I'd compile this for historical reasons.

http://www.mediafire.com/?i2gy1k0dmwm

Jim.

CHESS revision 5-23-88 by John Stanback
http://john.stanback.net/

Final program by John Stanback that contributed to
the GnuChess engine. First released 5-23-88.

Converted to Winboard protocol.

http://www.mediafire.com/?wlntm5ajttz

Jim.

Quick update -
Chess 5-23-88 Rev 0.2 by John Stanback

Fixed mate bug.
Fixed promotion bug.
Resign added.

http://www.mediafire.com/?zzmmmhmyjwm

Jim.

Hi Jim,

We now have 2 engines named "Chess":

Chess 5-23-88 Rev 0.2 by John Stanback
and
Chess 2.9a by Evgeniy Korniloff
http://evgeniy-korniloff.narod.ru/download.html

Very confusing!

Ron

Hi Ron,

Yes, but John Stanback's was first. By twenty years! First version of Chess was released 1988.
I name Evgeny Korniloff's engine 'Chess One' because that's what it's called in the source code.

Jim.

There seems to be a fair few bugs in this early version.

Another update. This fixes a bug in sending the mate result correctly.

Chess 5-23-88 Rev 0.3 by John Stanback
http://www.mediafire.com/?ofm2mntyjy2

Jim.

Another bug found .... can't be many left now - I hope!

This update fixes a resign bug. (Engine would resign immediately after it had made it's move, instead of before)

Chess 5-23-88 Rev 0.4 by John Stanback

http://www.mediafire.com/?jmjofyyjmgj

Jim.

Hi Jim

As usual, many many thanks for bringing these oldies back to life

I'll include both Chess and GnuChess 3.1 in my next Promo tournament.

Olivier

Thanks Olivier.

Both Chess 5-23-88 and GnuChess 3.1 are surprisingly quite strong
for their age. They must have been quite something when first released.
GnuChess 3.1 has ponder capability, which I've just discovered. My next
task will be to get it working correctly in the winboard version, it's not enabled at the moment.

best,
Jim.

Jim Ablett wrote:Hi Ron,

Yes, but John Stanback's was first. By twenty years! First version of Chess was released 1988.

I agree completely. No new author should use the name of an older engine!

We have the same problem with the name 'Chaturanga.' If the original Chaturanga source becomes available there will be another problem!

I wonder if there are other engines with duplicate names?

Jim Ablett wrote:I name Evgeny Korniloff's engine 'Chess One' because that's what it's called in the source code.

Jim.

When I run Evgeny Korniloff's engine in console mode, it prints "Chess" as the name. I don't have access to the source code, so I will change my information to 'Chess One' based on your observation. But maybe 'Chess Two' would be more accurate!

One of the reasons for the Private Engine List is to reserve names for future chess engines. The list also exposes naming conflicts in advance of a public release.

In my opinion: When there is a conflict in names, the original program is the one that gets to keep its name and the newer one using its name gets renamed!

Thanks, Jim!

Ron

Ron Murawski wrote:
I wonder if there are other engines with duplicate names?

Eden


.

Thanks Jim; i like GnuChess and these versions are of collection.
Maybe Chess 5-23-88 might be named GnuChess 5-23-88 or GnuChess 1.0.

Ron Murawski wrote:
One of the reasons for the Private Engine List is to reserve names for future chess engines. The list also exposes naming conflicts in advance of a public release.

...

Ron

Hi Ron, Brainless isn't private anymore.

Regards,
Guenther