Doch 1.3 : 2635

Discussion of computer chess matches and engine tournaments.

Moderator: Ras

User avatar
Le Fou numerique
Posts: 859
Joined: Wed Apr 05, 2006 9:58 pm
Location: Elsass

Doch 1.3 : 2635

Post by Le Fou numerique »

Hi,

ImageDoch 1.3 first rating in the UEL: 2635 (1.2: 2708, 404 games rated).

PGN on the page of the Division Elite.

Regards,
Patrick
DomLeste
Posts: 221
Joined: Thu Mar 09, 2006 4:53 pm

Re: Doch 1.3 : 2635

Post by DomLeste »

perhaps you can test Doch 1.33 as a comparison?
Insanity: doing the same thing over and over again and expecting different results.
Albert Einstein
User avatar
Graham Banks
Posts: 44641
Joined: Sun Feb 26, 2006 10:52 am
Location: Auckland, NZ

Re: Doch 1.3 : 2635

Post by Graham Banks »

DomLeste wrote:perhaps you can test Doch 1.33 as a comparison?
No strength difference. Just a bugfix for personalities.
gbanksnz at gmail.com
User avatar
Le Fou numerique
Posts: 859
Joined: Wed Apr 05, 2006 9:58 pm
Location: Elsass

Re: Doch 1.3 : 2635

Post by Le Fou numerique »

Hi Dom,

This tournament starts befor the new versions of Doch were out.

The default sample.per was used.

The tournament with 1.3.3 version will come.

Regards,
Patrick
DomLeste
Posts: 221
Joined: Thu Mar 09, 2006 4:53 pm

Re: Doch 1.3 : 2635

Post by DomLeste »

Doch 1.3x is strongest without the sample.per entry personality...will shall see :)
Insanity: doing the same thing over and over again and expecting different results.
Albert Einstein
swami
Posts: 6662
Joined: Thu Mar 09, 2006 4:21 am

Re: Doch 1.3 : 2635

Post by swami »

Le Fou numerique wrote:Hi,

ImageDoch 1.3 first rating in the UEL: 2635 (1.2: 2708, 404 games rated).

PGN on the page of the Division Elite.

Regards,
Patrick
Since Doch 1.3 was tested in STS and was proven to be weaker than Doch 1.2, this results confirms it.

Perhaps bug fixed 1.3.3 could do better, Will run the test.
Carlos777
Posts: 1944
Joined: Sun Dec 13, 2009 6:09 pm

Re: Doch 1.3 : 2635

Post by Carlos777 »

Le Fou numerique wrote: The default sample.per was used.
Hi Patrick,

I think Doch could be doing better without using the personality file.

According to Don Dailey:
The result is that if load a personality you will have a weaker program, even if you use the default.

If you don't use a personality, the program will play at full strength.

Don
Best,
Carlos
User avatar
Le Fou numerique
Posts: 859
Joined: Wed Apr 05, 2006 9:58 pm
Location: Elsass

Re: Doch 1.3 : 2635

Post by Le Fou numerique »

Hi,

I had no other possibility to run Doch 1.3. Without personnality file it crash.

I have some results with Doch 1.3.3 without personnality file. It seems better.

Regards,
Patrick
User avatar
Le Fou numerique
Posts: 859
Joined: Wed Apr 05, 2006 9:58 pm
Location: Elsass

Re: Doch 1.3 : 2635

Post by Le Fou numerique »

swami wrote:
Perhaps bug fixed 1.3.3 could do better, Will run the test.
Hi Swami,

Here some results with Doch 1.3 and Doch 1.3.3:

Code: Select all

5+2 64Mo HS-20moves.abk
                              Score     St Do De Sp Tw Cr Co Do Al Bo Cy Sp Bi
-------------------------------------------------------------------------------
 1: Stockfish 1.6.2         19.5 / 24   XX 11 01 11 =1 == 11 1= 11 11 =1 1= =1
 2: Doch 1.3.3              15.5 / 24   00 XX 1= 01 1= == =0 =1 11 =1 =1 =1 11
 3: Delfi 5.4               14.5 / 24   10 0= XX == == 0= 11 1= =1 =1 10 1= 1=
 4: Spark 0.3               13.0 / 24   00 10 == XX 01 1= 10 0= 1= 1= 0= 1= 11
 5: Twisted Logic 20090922  12.5 / 24   =0 0= == 10 XX 10 11 0= =1 1= == == 10
 6: Critter 0.42            12.0 / 24   == == 1= 0= 01 XX 1= == 00 11 00 11 0=
 7: Colossus 2008b          12.0 / 24   00 =1 00 01 00 0= XX 11 == 10 11 11 ==
 8: Doch 1.3                11.5 / 24   0= =0 0= 1= 1= == 00 XX 01 =1 11 =0 01
 9: Alaric 707              10.5 / 24   00 00 =0 0= =0 11 == 10 XX 00 11 == 11
10: Booot 4.15.0            10.0 / 24   00 =0 =0 0= 0= 00 01 =0 11 XX 11 =1 ==
11: Cyrano 0.6 b17           9.0 / 24   =0 =0 01 1= == 11 00 00 00 00 XX 10 1=
12: Spike 1.2 Turin          8.0 / 24   0= =0 0= 0= == 00 00 =1 == =0 01 XX 01
13: Bison 9.11               8.0 / 24   =0 00 0= 00 01 1= == 10 00 == 0= 10 XX
-------------------------------------------------------------------------------
156 games: +56 =55 -45
Regards,
Patrick
User avatar
beachknight
Posts: 3533
Joined: Tue Jan 09, 2007 8:33 pm
Location: Antalya, Turkey

Re: Doch 1.3 : 2635

Post by beachknight »

A replication with Doch x64 versions:

Code: Select all

    Engine                 Score StSpCyDoDoSpCrTwBiDeAlCoBo    S-B
01: Stockfish 1.6.0 DC     0.0/0 ·                            0,00
01: Spark 0.3 vc           0.0/0   ·                          0,00
01: Cyrano 0.6b17 JA       0.0/0     ·                        0,00
01: Doch 1.3.3 JA          0.0/0       ·                      0,00
01: Doch 1.3 JA            0.0/0         ·                    0,00
01: Spike 1.2 w32          0.0/0           ·                  0,00
01: Critter 0.42 w32       0.0/0             ·                0,00
01: Twisted Logic 20090922 0.0/0               ·              0,00
01: Bison 9.11             0.0/0                 ·            0,00
01: Delfi 5.4 1cpu w32     0.0/0                   ·          0,00
01: Alaric 0707 w32        0.0/0                     ·        0,00
01: Colossus 2008b w32     0.0/0                       ·      0,00
01: Booot 4.15.0 w32       0.0/0                         ·    0,00

0 of 156 games played
Name of the tournament: CET2403p
Site/ Country: CET24_Antalya_TUR, Türkiye
Level: Blitz 1/1
hi, merhaba, hallo HT