WBEC Ridderkerk new results.

Archive of the old Parsimony forum. Some messages couldn't be restored. Limitations: Search for authors does not work, Parsimony specific formats do not work, threaded view does not work properly. Posting is disabled.

WBEC Ridderkerk new results.

Postby Leo Dijksman » 27 Jul 2004, 15:40

Geschrieben von:/Posted by: Leo Dijksman at 27 July 2004 16:40:05:

1st Division:
Round 9:
Movei 0.08.247 0½11 2.5/4
Abrok 5.0 1½00 1.5/4
---------------------------------
Maestro 1.06uci 1½½½ 2.5/4
Comet B.68 0½½½ 1.5/4
---------------------------------
Round 10:
WildCat 4.0 ½1½½ 2.5/4
AnMon 5.32 ½0½½ 1.5/4
---------------------------------

3rd Division:
Round 10:
Resp 0.19 1111 4.0/4 !!
Cerebro 1.26g 0000 0.0/4
---------------------------------
Averno 0.70a 1½10 2.5/4
Alarm 0.93.1 0½01 1.5/4
---------------------------------
DanChess 1.06d 0111 3.0/4
Djinn 0.870f 1000 1.0/4
---------------------------------

5th Division:
Round 16:
Nesik 0.6.4 - Gerbil 0.2 =-= 89
Round 17:
LaDameBlanche 2.0c - Delphil 0.8e 1-0 45
Delphil 0.8e - LaDameBlanche 2.0c =-= 110
Ax 0.8 - Eeyore 1.39 =-= 127
Eeyore 1.39 - Ax 0.8 1-0 62
Tornado 0.80.17b - Gully 2.15 pl10 1-0 51
Gully 2.15 pl10 - Tornado 0.80.17b 0-1 73
Sharper 0.17 - SseChess 2.04.5 1-0 36
SseChess 2.04.5 - Sharper 0.17 1-0 80
DrunkenMaster 1.1 - Matheus 2.3 0-1 74
Matheus 2.3 - DrunkenMaster 1.1 0-1 101
--------------------------------------------------------------

6th Division Group C:
Round 4:
Grizzly 1.40.1b - Borer 2.01.8a 0-1 88
Damas 6a3 - Cilian 4.14 =-= 52
Cilian 4.14 - Damas 6a3 1-0 40
SdBC 0.4.13.0 - Gargamella 0.7.0 1-0 55
Gargamella 0.7.0 - SdBC 0.4.13.0 1-0 48
Enigma 1.1.4 - TSCP 1.81 =-= 143
TSCP 1.81 - Enigma 1.1.4 1-0 58
Geko 0.4.3 - Adamant 1.5 0-1 12
Adamant 1.5 - Geko 0.4.3 1-0 30
Round 5:
Grizzly 1.40.1b - AICE 0.72.3 0-1 51
AICE 0.72.3 - Grizzly 1.40.1b 1-0 53
SnailChess 4.013 - Pooky 2.3 1-0 28
Pooky 2.3 - SnailChess 4.013 =-= 53
Cilian 4.14 - Borer 2.01.8a 0-1 53
Borer 2.01.8a - Cilian 4.14 1-0 50
--------------------------------------------------------------

Crosstables and pgn on the WBEC homepage.
Enginelist, latest updates/new engines:
27/07/2004:
Updated: Patriot 1.3.0 (Commercial, free update for v1.1.0 owners! also UCI!)
Updated: FireFly v1.1.1 (also UCI!)
Updated: KKFChess v2.6.6 (also UCI)
Updated: BigLion v2.23m (also UCI!)
Updated: Taktix v2.23m (also UCI!)
Updated: Alex v1.60.09e (also UCI!)
Best wishes,
Leo.


WBEC Ridderkerk homepage.
Leo Dijksman
 

Re: WBEC Ridderkerk new results.

Postby Uri Blass » 27 Jul 2004, 17:03

Geschrieben von:/Posted by: Uri Blass at 27 July 2004 18:03:41:
Als Antwort auf:/In reply to: WBEC Ridderkerk new results. geschrieben von:/posted by: Leo Dijksman at 27 July 2004 16:40:05:
1st Division:
Round 9:
Movei 0.08.247 0½11 2.5/4
Abrok 5.0 1½00 1.5/4
---------------------------------

Movei was not supposed to win drawn pawn endgames and the best that I expected from it is to draw if it is lucky not to blunder.
It does not use hash tables for pruning and it should be a big problem in pawn endgame.
It also does not use tablebases
It also has not a lot of knowledge for pawn endgame but for some unclear reason Abrok blundered by g4 so movei played g5 and won the game
relevant positions
8/8/4k1p1/1p2p2p/1P2K2P/p4PP1/P7/8 w - b6 0 49 am g4
8/8/4k1p1/1p2p2p/1P2K1PP/p4P2/P7/8 b - - 0 49 bm g5
Movei that I considered to be one of the weakest programs in pawn endgames solve both positions easily.
Uri
Uri Blass
 

Re: WBEC Ridderkerk new results.

Postby Luis Smith » 27 Jul 2004, 18:43

Geschrieben von:/Posted by: Luis Smith at 27 July 2004 19:43:41:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Uri Blass at 27 July 2004 18:03:41:

Hello Uri,
I see Movei is doing quite well. What changes did you do from .232 to 247 to get a similar performance to WildCat?


Rank:     Engine:                  Rating Rating Games
New: Diff:                         New:   Diff:  Total:
27   1    WildCat 3.07             2501   24     512
40   -1   Movei 0.0.8.232          2434   19     458


Also note that if Leo's 1st division tournament stopped right now that Movei has very good chances to promote. El Chinito looks as if it is already going to promote with 2nd 3rd and 4th look like this:


2: WildCat 4.0          23.5 / 40
3: Movei 0.08.247       22.0 / 36
4: Maestro 1.06uci      20.5 / 36   


In heads up matches Movei has beaten both engines:


Movei 0.08.247       11½½  3.0/4 
WildCat 4.0          00½½  1.0/4
----------------------------------
Movei 0.08.247       1½½1  3.0/4 
Maestro 1.06uci      0½½0  1.0/4
Luis Smith
 

Re: WBEC Ridderkerk new results.

Postby Uri Blass » 27 Jul 2004, 19:16

Geschrieben von:/Posted by: Uri Blass at 27 July 2004 20:16:29:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Luis Smith at 27 July 2004 19:43:41:
Hello Uri,
I see Movei is doing quite well. What changes did you do from .232 to 247 to get a similar performance to WildCat?



Also note that if Leo's 1st division tournament stopped right now that Movei has very good chances to promote. El Chinito looks as if it is already going to promote with 2nd 3rd and 4th look like this:



In heads up matches Movei has beaten both engines:



I do not know.
I did changes in the search and the evaluation and the learning but I guess that the most important thing that I added was adding some luck.
I still do not expect it to promote because I am afraid that I did not add enough luck.
one change that I did in the learning was to have more aggresive learning and movei tries to repeat lines that it won so it always repeat the 1.e4 c6 line with black after it won with it.
I am afraid that it may cost it points because losing a game reduce the learning score only by 1 so after many wins it needs many losses in the same line in order not to repeat it but fortunately a lot of opponents here have not aggresive learning and do not try to repeat wins so until now it never repeated lost games.
I did not expect it to score so well and I prefered to have also positive learning and not only negative learning because my assumption was that the stupid thing is not going to do well so if there is a line when it scores more than 50% it is better to repeat it even if it lost the last game with that line because the possible advantage from repeating a good line is bigger than the risk from repeating the same loss again.
There is also hope that it will not repeat the same loss again even if the same opening repeats again because if the pondering time is not exactly the same it may have different order of pieces and it may cause different moves later.
>Rank:     Engine:                  Rating Rating Games
>New: Diff:                         New:   Diff:  Total:
>27   1    WildCat 3.07             2501   24     512
>40   -1   Movei 0.0.8.232          2434   19     458
>
>2: WildCat 4.0          23.5 / 40
>3: Movei 0.08.247       22.0 / 36
>4: Maestro 1.06uci      20.5 / 36   
>
>Movei 0.08.247       11½½  3.0/4 
>WildCat 4.0          00½½  1.0/4
>----------------------------------
>Movei 0.08.247       1½½1  3.0/4 
>Maestro 1.06uci      0½½0  1.0/4
>
Yes but I expect them to learn and not repeat the same mistakes again and I expect Maestro to be replaced by Falcon that is clearly better.
Maestro is an old version of Falcon before graz.
Uri
Uri Blass
 


Return to Archive (Old Parsimony Forum)

Who is online

Users browsing this forum: No registered users and 18 guests