FEOBOS v20.1 Contempt = 3, 40 in 40 test-tourney, 22 eng.

[jdart]

Thanks for the PGN link.

--Jon

[Kotlov]

How about new project after that?

[Frank Quisinsky]

Hi Eugene,

works great.

Once is A00-A99 ... over 20% ... 5-6% to high if I compare with GM theory. All other things works. 1 game with draw under 20 moves after 964 games (a short line, impossible to avoid all of such lines).

The question is:
Should I set priorities by hand or not in the final FEOBOS version. Normally 20% A00-A99 is right because more lines possible in comparing to C00-C99. So stats are right but in GM theory "with move transpositions" we have lesser A00-A99 and more lines with move transpositions to C00-C99.

After all that what I can see for FEOBOS 20.1 Contempt 3 ... again it works great.

Best and thanks for your question.
Let us wait of the final results from test-2 tourney with Top-21 + Wasp. Must thinking what I do here ... "Priorities set by hand in final FEOBOS Version with Contempt = 3 or not".

Best
Frank

[Frank Quisinsky]

ECO A00-E99 isn't good.
Please have a look on FEOBOS page ...

Clemens, you have to set book options (can be found on the end of the page for Shredder and Fritz GUI). Graphics with book options can be found in download file to version 20.1 (very important to set the right book options).

Best
Frank

http://www.amateurschach.de/main/_new-opening-book.htm (end of the page).

From my tourney with the settings you can found on the end of the page I have after 964 games ...

Contempt = 3, FEOBOS 20.1 without move transpositions.

Code: Select all

Games        : 924
Move average : 89,86 (without resign)

White Wins   : 274 (29.7%)
Black Wins   : 183 (19.8%)
Draws        : 467 (50.5%)
Unfinished   : 0

White Perf.  : 54.9%
Black Perf.  : 45.1%

ECO A = 199 Games (21.5%)
ECO B = 268 Games (29.0%)
ECO C = 142 Games (15.4%)
ECO D = 163 Games (17.6%)
ECO E = 152 Games (16.5%)

[Frank Quisinsky]

Hi there,

the second FEOBOS 20.1 Contempt 3 test tourney is over. Book statistics are best I ever saw. So the 10 FEOBOS engines do a great work in teamwork.

With 66,0 / 84 points Komodo 11.3.1 on first.
With 65,5 / 84 points Houdini 6.03 on second.
With 65,0 / 84 points Stockfish 9 on third.

The final results from the 40 / 40 tourney on i7 4.3 Ghz can be found with PGN download and stats directly on my start page.

Have fun with it.

Tomorrow I start FEOBOS 20.1 Contempt 3 test-tourney 3.



Frank's Chess Page
http://www.amateurschach.de

Best
Frank

[Frank Quisinsky]

Hi there,

now the third test.
Again a group with 22 engines.
We can say the second best group of available engines, places 20-40 in rating systems like CCRL / CEGT.

Link can be found on my website.

Have fun

Best
Frank

[Frank Quisinsky]

Hi there,

I have here much Problems with Bobcat 8 (without resign, end of the game, GUI Crash before game ended in around 10% of cases).

So I added the derivat Tomcat 1.0 in the list of engines.
Ethereal will play now the second bigger tourney and the first for Xiphos.

2.950 - 2.750 from participant 1 - 22 ... will be very interesting.

What I can see in the first two tourneys:

1. Fritz 16 have very big problems with tactical positions and many pieces on board, lost many games very fast in the second tourney.

2. ICE is maybe much stronger with longer time controls (I am using here 40 moves in 30 minutes repeatedly on i7 4.3Ghz systems.

3. It seems that Houdini made more poins vs. weaker opponents as Komodo and Stockfish, but all have around the same strength and playing style.

Best
Frank

[Frank Quisinsky]

Hi there,

no changes are necessary.
FEOBOS v20.1 Contempt 3 works pefect.

The final results from the three test-tourneys can be found with interesting statistics on my start page.

Frank's Chess Page
http://www.amateurschach.de

Best
Frank