WBEC Ridderkerk new results.

[Leo Dijksman]

Geschrieben von: / Posted by: Leo Dijksman at 16 March 2004 16:50:02:

1st Division:
Round 11:
SOS 99.11.03 1½11 3.5/4
King Of Kings 2.52 0½00 0.5/4
------------------------------------
Comet B.68 101½ 2.5/4
XiniX 04.02.09 010½ 1.5/4
------------------------------------
Round 12:
El Chinito 3.25 111½ 3.5/4
Patzer 3.61 000½ 0.5/4
------------------------------------
Abrok 5.0 011 2.0/3 = 1 game to go!
Leila 0.53h 100 1.0/3
------------------------------------

3rd Division:
Round 14:
Maestro Exp. 1110 3.0/4
Dorky 3.48 0001 1.0/4
------------------------------------
Queen 2.42a 1½½1 3.0/4
TRACE 1.25 0½½0 1.0/4
------------------------------------
Averno 0.70a 1110 3.0/4
Ant 6.06 0001 1.0/4
------------------------------------
Horizon 4.1b1 111 3.0/3 = 1 game to go!
Gaviota 0.33 000 0.0/3
------------------------------------

5th Division: Finished!
Last round:
Cilian 4.13 - Mustang 2.67 0-1 125 = No result or debugfile found, adjudicated to a win for Mustang!
DrunkenMaster 0.9 - EnginMax 5.11c 1-0 87
EnginMax 5.11c - DrunkenMaster 0.9 0-1 47 = Good tourney result by DrunkenMaster, played the biggest part with ponder=off!!
RDChess 3.23 - Qalat 0.19b 0-1 75
Qalat 0.19b - RDChess 3.23 1-0 47
Ranita 2.4 - Ax 0.8 0-1 28
Ax 0.8 - Ranita 2.4 1-0 43
Grizzly 1.40.1b - Faile 1.4 1-0 84
Faile 1.4 - Grizzly 1.40.1b 1-0 77 = Grizzly says {technical draw!!!} in a hopeless position, set to 1-0!
Rinko 1.15 - Waster 0.13 0-1 48
Waster 0.13 - Rinko 1.15 1-0 61
Bodo 0.2e - MrChess 2.1 1-0 50 = Bodo qualified for he promotion tourney!
MrChess 2.1 - Bodo 0.2e 1-0 34
-------------------------------------------------------------

6th Division:
Round 19:
Robin 0.9.86 - Awesome 1.45 1-0 60
Rainman 0.7.5 - Adamant 1.5 1-0 33
Smash 0.8d - ChessRikus 1.4.60 0-1 82
Parrot 031231 - Matacz 0.93 1-0 82
Cefap 0.7.2 - SdBC 0.4.13.0 0-1 58
Golem 0.4 - Beaches 1.52 =-= 128
Ozwald 0.43 - Mint 2.3 0-1 87
Geko 0.4.3 - ApiChess 1.29 0-1 47
Damas 6a2 - Chad'sChess 1.5 1-0 29
Bace 0.46 - Yawce 0.16 0-1 34
MiniChessAI 1.18 - LarsenVB 0.05.01 =-= 58
KillerQueen 1.0.1 - StAndersen 1.31 =-= 49
Dimitri 0.6.1 - Fafis 0.9 0-1 33
WESP 0.8 - Nero 5.3 1-0 41
Pooky 2.1 - Aice 0.55 0-1 89
-------------------------------------------------------------

Crosstables and pgn on the WBEC homepage.
Enginelist, latest updates/new engines:
16/03/2004:
New Engine: Elephant v1.00 (Be carefull, use around 150MB of your RAM!)

Leo.


WBEC Ridderkerk homepage.

[Ubaldo Andrea Farina]

Geschrieben von: / Posted by: Ubaldo Andrea Farina at 16 March 2004 17:57:40:
Als Antwort auf:/In reply to: WBEC Ridderkerk new results. geschrieben von: / posted by: Leo Dijksman at 16 March 2004 16:50:02:

Hi all,
Leo thanks for this great tourney. Congratulations to Waster and good luck to Bodo, NagaSkaki and (of course ) Chiron for the promotion tourney!
--Ubaldo

[Ubaldo Andrea Farina]

Geschrieben von: / Posted by: Ubaldo Andrea Farina at 16 March 2004 18:01:54:
Als Antwort auf:/In reply to: Re: My God, Volker, what have you done!? (n/t) geschrieben von: / posted by: Ubaldo Andrea Farina at 16 March 2004 17:57:40:

nt

[Marcus Prewarski]

Geschrieben von: / Posted by: Marcus Prewarski at 16 March 2004 18:39:10:
Als Antwort auf:/In reply to: WBEC Ridderkerk new results. geschrieben von: / posted by: Leo Dijksman at 16 March 2004 16:50:02:

EnginMax 5.11c - DrunkenMaster 0.9 0-1 47 = Good tourney result by DrunkenMaster, played the biggest part with ponder=off!!

Thanks Leo, I'm happy with the result. Congratulations to all the engines advancing, hopefully 5th division will be a more friendly place without them . Although with new engines and improving old ones it will get tougher each edition. I think there is a bit of overlap between the divisions and the top ones should do fine in division 4.
For anyone interested, pondering is a fairly new feature for DrunkenMaster and I had bugs with time management while pondering (losses on time) so I should have these worked out for next edition.
In about a month I will release a new version with:
- a UCI interface (more potential for bugs)
- fixed time management problems.
- a completely rewritten search. My current one is still buggy and I hope
to fix most of the problems and I think this will improve the engine a bit.
My longer term goal is to follow the path of Gothmog and I believe Smarthink and have the search tied to the evaluation but before doing this I felt I needed to get the basics worked out and more solid. I haven't seen the source code for Fruit, but judging from the authors comments, it is a fairly simple engine yet plays quite well.
-Marcus

[Matthias Gemuh]

Geschrieben von: / Posted by: Matthias Gemuh at 16 March 2004 19:20:33:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Marcus Prewarski at 16 March 2004 18:39:10:

- a UCI interface

My longer term goal is to follow the path of Gothmog ...

Hi Marcus,
BigLion is going same way.
As work-around for the Fritz GUI 1 MB Hash issue, Gothmog UCI has fritz-hash option, while BigLion UCI would use hash size from INI file.
Do something similar, so that I can use DrunkenMaster UCI as sparring partner
for BigLion UCI in Fritz GUI.
/Matthias.



BigLion + Taktix

[Tord Romstad]

Geschrieben von: / Posted by: Tord Romstad at 16 March 2004 19:41:37:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Marcus Prewarski at 16 March 2004 18:39:10:

Hi Marcus!

EnginMax 5.11c - DrunkenMaster 0.9 0-1 47 = Good tourney result by DrunkenMaster, played the biggest part with ponder=off!!

Thanks Leo, I'm happy with the result. Congratulations to all the engines advancing, hopefully 5th division will be a more friendly place without them . Although with new engines and improving old ones it will get tougher each edition.
I think there is a bit of overlap between the divisions and the top ones should do fine in division 4.
For anyone interested, pondering is a fairly new feature for DrunkenMaster and I had bugs with time management while pondering (losses on time) so I should have these worked out for next edition.
In about a month I will release a new version with:
- a UCI interface (more potential for bugs)
- fixed time management problems.
- a completely rewritten search. My current one is still buggy and I hope
to fix most of the problems and I think this will improve the engine a bit.
My longer term goal is to follow the path of Gothmog and I believe Smarthink and have the search tied to the evaluation but before doing this I felt I needed to get the basics worked out and more solid. I haven't seen the source code for Fruit, but judging from the authors comments, it is a fairly simple engine yet plays quite well.

DrunkenMaster seem to be among the engines which improve most quickly at the
moment, and if progress continues at the current rate I don't think you have
any reason to fear the next edition.
Good idea. There are a couple of UCI-only tournaments which by themselves
are sufficient reason that adding UCI support is worth the effort, IMHO.
As Matthias has already mentioned, it's a good idea to include some kind
of workaround for the 1MB hash bug in the Fritz GUI.
Perhaps you have seen it already, but I wrote a description of how my
search works on the CCC yesterday. You will find it in the thread named
"Teaching computers to think".
I agree that it's a good idea to have something simple, stable and
bug-free before you start adding more complicated stuff.
Good luck with the next version! DrunkenMaster 0.9 is already a very
interesting program, and one of my favorites among the Linux engines.
Tord

[Marcus Prewarski]

Geschrieben von: / Posted by: Marcus Prewarski at 16 March 2004 19:54:54:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Matthias Gemuh at 16 March 2004 19:20:33:

Hi Marcus,
BigLion is going same way.
As work-around for the Fritz GUI 1 MB Hash issue, Gothmog UCI has fritz-hash option, while BigLion UCI would use hash size from INI file.
Do something similar, so that I can use DrunkenMaster UCI as sparring partner
for BigLion UCI in Fritz GUI.
/Matthias.

I haven't seen the bug myself but I've read about it from you and Tord and right now I'm handling it the same way you are. Currently I'm using Arena to debug my UCI support, how do you debug with the Fritz interface, write all the input/output to a file yourself? Now that I've added UCI support I can say that I think it more simple to implement then WinBoard, especially pondering. But maybe I will run into problems that are hard to solve with UCI.
Hopefully DrunkenMaster will make a good sparring parnter, just yesterday I downloaded the latest Big Lion and plan to use it as well. Good luck with your engine.
-Marcus

[Marcus Prewarski]

Geschrieben von: / Posted by: Marcus Prewarski at 16 March 2004 20:19:58:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Tord Romstad at 16 March 2004 19:41:37:

DrunkenMaster seem to be among the engines which improve most quickly at the
moment, and if progress continues at the current rate I don't think you have
any reason to fear the next edition.
Good idea. There are a couple of UCI-only tournaments which by themselves
are sufficient reason that adding UCI support is worth the effort, IMHO.
As Matthias has already mentioned, it's a good idea to include some kind
of workaround for the 1MB hash bug in the Fritz GUI.
Perhaps you have seen it already, but I wrote a description of how my
search works on the CCC yesterday. You will find it in the thread named
"Teaching computers to think".
I agree that it's a good idea to have something simple, stable and
bug-free before you start adding more complicated stuff.
Good luck with the next version! DrunkenMaster 0.9 is already a very
interesting program, and one of my favorites among the Linux engines.
Tord

The key to rapid improvement is to start with a lot of bugs . The largest improvements lately have come from debugging my search. But a few bugs still remain and so I decided to completely rewrite it, the code was getting more and more ugly too. In my development code I've switched from PVS to negascout and plan to experiment with MTD. One of the reasons I switched is I wanted a fresh approach so I would be less tempted to cut and paste my old code.

Yes I've seen the thread and even saved a copy, it is very interesting stuff. I really can't wait to start working on some of these ideas, but I'm just not ready yet. I really appreciate how open you are about design and sharing your ideas. And I think it is much better than providing open source code.
Thanks for the kind words, now that I have access to a linux machine I plan to continue to provide linux versions.
-Marcus

[Matthias Gemuh]

Geschrieben von: / Posted by: Matthias Gemuh at 16 March 2004 21:24:38:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Marcus Prewarski at 16 March 2004 19:54:54:

Currently I'm using Arena to debug my UCI support, how do you debug with the >Fritz interface, write all the input/output to a file yourself?


just yesterday I downloaded the latest BigLion and plan to use it as well.

Yes.

2.23k ? Bad idea. 2.23i is much stronger.
/Matthias.


BigLion + Taktix

[Uri Blass]

Geschrieben von: / Posted by: Uri Blass at 16 March 2004 21:46:18:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Marcus Prewarski at 16 March 2004 20:19:58:

DrunkenMaster seem to be among the engines which improve most quickly at the
moment, and if progress continues at the current rate I don't think you have
any reason to fear the next edition.
Good idea. There are a couple of UCI-only tournaments which by themselves
are sufficient reason that adding UCI support is worth the effort, IMHO.
As Matthias has already mentioned, it's a good idea to include some kind
of workaround for the 1MB hash bug in the Fritz GUI.
Perhaps you have seen it already, but I wrote a description of how my
search works on the CCC yesterday. You will find it in the thread named
"Teaching computers to think".

The key to rapid improvement is to start with a lot of bugs . The largest improvements lately have come from debugging my search. But a few bugs still remain and so I decided to completely rewrite it, the code was getting more and more ugly too. In my development code I've switched from PVS to negascout and plan to experiment with MTD. One of the reasons I switched is I wanted a fresh approach so I would be less tempted to cut and paste my old code.

Yes I've seen the thread and even saved a copy, it is very interesting stuff. I really can't wait to start working on some of these ideas, but I'm just not ready yet.

same for me.
I am also not ready because I first need to implement some basic stuff.
Uri