Page 30 of 39

Re: Never tested by both CCRL & CEGT (GopherCheck w32)

Posted: Wed Mar 09, 2016 1:29 am
by carldaman
Where did you download the Gopher_Check executable, Herr Ruxy?

I tried using Daniel Jose's 64-bit binary, but it refuses to make a move in Arena.

Thanks,
CL

Sylwy wrote:Hello !

Conditions:

-TC = 6'/move (the first match); 8'/move (the tournament)
-Hash = 128 MB
-Book = Perfect_2010.abk
-GUI = Arena 3.0
-CPU = Athlon 64 3800+

Image

Image

My verdict for GopherCheck w32 (a chess engine written in Go) = very close to 1870 Elo points (CCRL 4/40).

Cheers,
SilvianR :wink:

Re: Never tested by both CCRL & CEGT (GopherCheck w32)

Posted: Wed Mar 09, 2016 9:20 am
by Sylwy
carldaman wrote:Where did you download the Gopher_Check executable, Herr Ruxy?

I tried using Daniel Jose's 64-bit binary, but it refuses to make a move in Arena.

Thanks,
CL
Hello !

I must recognize that's a great drama ! :lol:

Serenissimo (near 1,90 m) Ruxy (I renounced to my unexisting Teutonic roots.... :lol: ) knows what's about with this naughty chess engine!

Use (PLEASE !!!) only TC= x seconds/move (for example 8 seconds/move) ! That's all ! This engine graduated only the kindergarten of computer chess.........


Have the supreme day of your existence !!!

Re: Never tested by both CCRL & CEGT (GopherCheck w32)

Posted: Fri Mar 11, 2016 3:55 am
by carldaman
Well, thanks again, Grandissimo Ruxy, wouldn't you know it, you were right again! ;) I was able to use Gopher-Check_64-bit as you suggested.

Great to count on your help. :)

Teutonic (maybe?) regards,
CL 8-)

Sylwy wrote:
carldaman wrote:Where did you download the Gopher_Check executable, Herr Ruxy?

I tried using Daniel Jose's 64-bit binary, but it refuses to make a move in Arena.

Thanks,
CL
Hello !

I must recognize that's a great drama ! :lol:

Serenissimo (near 1,90 m) Ruxy (I renounced to my unexisting Teutonic roots.... :lol: ) knows what's about with this naughty chess engine!

Use (PLEASE !!!) only TC= x seconds/move (for example 8 seconds/move) ! That's all ! This engine graduated only the kindergarten of computer chess.........


Have the supreme day of your existence !!!

Re: Never tested by both CCRL & CEGT (GopherCheck w32)

Posted: Fri Mar 11, 2016 9:08 am
by Sylwy
carldaman wrote:Well, thanks again, Grandissimo Ruxy, wouldn't you know it, you were right again! ;) I was able to use Gopher-Check_64-bit as you suggested.

Great to count on your help. :)

Teutonic (maybe?) regards,
CL 8-)

Sylwy wrote:
carldaman wrote:Where did you download the Gopher_Check executable, Herr Ruxy?

I tried using Daniel Jose's 64-bit binary, but it refuses to make a move in Arena.

Thanks,
CL
Hello !

I must recognize that's a great drama ! :lol:

Serenissimo (near 1,90 m) Ruxy (I renounced to my unexisting Teutonic roots.... :lol: ) knows what's about with this naughty chess engine!

Use (PLEASE !!!) only TC= x seconds/move (for example 8 seconds/move) ! That's all ! This engine graduated only the kindergarten of computer chess.........


Have the supreme day of your existence !!!

:wink: I wish all the best to you ! :wink:

Re: Galjoen 0.32 w32 - my first test

Posted: Mon Mar 14, 2016 10:41 am
by Werner Taelemans
Sylwy wrote:-I used the engine like Winboard (like UCI stopped);
-I used a GUI book (Perfect_2010.abk) - I had the impression that Galjoen can't recognize (like Wb engine) its own .bin book;
-I used the TC = 5'/40 moves (with repetition). At my "classical" TC= 4'+2" were some loses on time.
Hello Ruxy,
I installed Arena 3.0 on my computer to test your findings, and I found:
1) Galjoen works as UCI engine, but you have to install it manually: Engines/Manage.../Details/Type -> UCI (Autodetect doesn't work).
2) Galjoen recognizes it's own book, but in Winboard mode Arena sends the command "OwnBook=true" instead of the Winboard command "OwnBook=1".
3) Indeed there was a bug in the timemanager (fixed now).

Re: Cheese 1.8 w32 test

Posted: Thu Mar 24, 2016 7:26 pm
by Sylwy
Conditions:

-TC = 4'+2"
-Hash = 128 MB
-Books = Shredder12.ctg
-GUI = Fritz 11
-CPU = Athlon 64 3800+
-OS = Windows XP SP3

Pharaon 3.5.1 was linked to Nalimov TBs.

Cheese 1.8 w32 has around 2625 Elo (CCRL 4/40).

Image

Image

Re: SmarThink 1.90 w32 comparative test

Posted: Thu Mar 31, 2016 7:30 pm
by Sylwy
-TC = 4'+2"
-Hash = 128 MB
-Book = Shredder12.ctg
-GUI = Fritz 11
-CPU = Athlon 64 3800+

SmarThink 1.90 was linked to Nalimov TBs.

Around +40 Elo points (CCRL 4/40) for the new SmarThink 1.90 w32. A superb chess engine !

Image

Image

Image

Re: Floyd 0.8 w32 comparative test

Posted: Mon Apr 04, 2016 6:50 pm
by Sylwy
Hello !

Conditions:

-TC = 4'+2"
-Hash = 128 MB
-Books = Perfect_2010.abk (Floyd 0.8) & its own book for Nemeton 1.3
-GUI = Arena 3.0
-CPU = Athlon 64 3800+
-OS = Windows XP SP3

Around 2410 Elo points (CCRL 4/40) for the new 32-bit version (+200 Elo face to the previous version.....WOW !).


A question (if possible) to Mr.van Kervinck: how strong (in blitz 4/40) is the last version of Rookie ? Will be nice to know !


Nemeton 1.3 has 2313 Elo CCRL 4/40.

Image

Image

Re: Floyd 0.8 w32 comparative test

Posted: Mon Apr 04, 2016 9:47 pm
by mvk
Sylwy wrote:A question (if possible) to Mr.van Kervinck: how strong (in blitz 4/40) is the last version of Rookie ? Will be nice to know !
I'm not recently making any changes to Rookie because it has become hard to change. When I made the v3 series I focussed on speed and weight tuning, but now I don't care much about those two as those are solved problems. Also work is quite busy and I travel a lot. Whenever I find time for toying around with machine learning it is easier to use Floyd because it is cleaner. I don't know how each will evolve. Maybe they will merge. Maybe both will stop and I will do another new one. Hard to tell.

Thanks for testing! I hope the games are interesting to observe. I normally only run it on the raspberry pi...

Re: Floyd 0.8 w32 comparative test

Posted: Tue Apr 05, 2016 10:02 am
by Sylwy
mvk wrote:
Sylwy wrote:A question (if possible) to Mr.van Kervinck: how strong (in blitz 4/40) is the last version of Rookie ? Will be nice to know !
I'm not recently making any changes to Rookie because it has become hard to change. When I made the v3 series I focussed on speed and weight tuning, but now I don't care much about those two as those are solved problems. Also work is quite busy and I travel a lot. Whenever I find time for toying around with machine learning it is easier to use Floyd because it is cleaner. I don't know how each will evolve. Maybe they will merge. Maybe both will stop and I will do another new one. Hard to tell.

Thanks for testing! I hope the games are interesting to observe. I normally only run it on the raspberry pi...

Yes ! Floyd 0.8 has a nice play. I watched a lot of games !
THANK YOU !