Page 39 of 40
Re: opening books are completely harmless
Posted: Wed Apr 02, 2025 2:02 pm
by deeds
------------
Notice:
This post in this topic has been reviewed and edited because it violated the OpenChess Forum Rules. This is a gentle reminder not to continue flaming this topic or others in the Forum. Further recurrent violations of the OpenChess Forum Rules will not be tolerated.
Thanks,
Best regards, Moderation.
------------
Already answered
here.
I use until 4 different learning engines as opponents since my 2nd learning session (15th learning session in progress...).
If the BIN book format hasn't a learning feature

not my problem !
If some engines haven't a learning feature

not my problem !
Re: opening books are completely harmless
Posted: Wed Apr 02, 2025 2:03 pm
by deeds
------------
Notice:
This post in this topic has been reviewed and edited because it violated the OpenChess Forum Rules. This is a gentle reminder not to continue flaming this topic or others in the Forum. Further recurrent violations of the OpenChess Forum Rules will not be tolerated.
Thanks,
Best regards, Moderation.
------------
Already answered
here.
No need another GUI to train learning engines and I already coded all the tools i needed for EXP files.
Re: opening books are completely harmless
Posted: Wed Apr 02, 2025 2:05 pm
by deeds
Sedat wrote:
...Baron 3.0 book, it's so weak...
Not my problem !
They are a few BIN books which contains ECO B lines :

So this book is used to train my EXP files.
Re: opening books are completely harmless
Posted: Wed Apr 02, 2025 2:30 pm
by Sedat Canbaz
@to Deeds
Sorry, but I disagree with you with
Almost all.. what you wrote again!
And once more,
In your recent Deeds 2025 games openings,
You are missing almost all of strong lines!
But you are free, no any problem..but
Reality is reality...
And as clear proof about over:
That's why your used openings performed
Too badly..only via 80+ Elo edge engine,
They started to be balanced e.g via 50+%
Win. percentage !
If still not clear,
You can continue to work over 1980 models,
But remember that exc these old models, our
World contains much newer, better MODELS!!
That's ALL over... and hope HELPS!
Re: opening books are completely harmless
Posted: Wed Apr 02, 2025 2:34 pm
by deeds
------------
Notice:
This post in this topic has been reviewed and edited because it violated the OpenChess Forum Rules. This is a gentle reminder not to continue flaming this topic or others in the Forum. Further recurrent violations of the OpenChess Forum Rules will not be tolerated.
Thanks,
Best regards, Moderation.
------------
And i also shared
tons of benchmark's results.
If the NPS displayed in the GUI are not the same as those saved in the PGN file

it's not my problem !
Re: opening books are completely harmless
Posted: Wed Apr 02, 2025 2:38 pm
by deeds
Sedat wrote:
...In your recent Deeds 2025 games openings, You are missing almost all of strong lines!...
No matter what you expect.
I train several learning engines on different opening lists to cover a wide variety of openings.
Since my EXP file's accuracy improves with each new opening learned, everything is fine.
Re: Deep Rybka 4.1 vs. its own playing style
Posted: Thu Apr 03, 2025 1:48 pm
by deeds
Next round in progress...

Re: opening books are completely harmless
Posted: Sat Apr 05, 2025 7:52 am
by ZamChess
All the best Sedat !
Re: opening books are completely harmless
Posted: Sat Apr 05, 2025 11:04 am
by Sedat Canbaz
ZamChess wrote: Sat Apr 05, 2025 7:52 am
All the best Sedat !
Thank you dear Zam

Re: Deep Rybka 4.1 vs. its own playing style
Posted: Tue Apr 08, 2025 10:38 am
by deeds
hardware : dual xeon e5-2660v3, 40 threads, hash 4 GB
software : Arena 3.5.1, TC 1m+1s, partial syzygy 7men
engine : Deep Rybka 4.1 played with own internal net
experience data : Deep_Rybka_4_1.exp (Experience Read Only=true, Experience Min Count=2, Experience Min Depth=16)
Opening list : lost_replay.pgn (the 181 non-drawish openings from round 1)
Code: Select all
# PLAYER : RATING ERROR POINTS PLAYED (%) W D L D(%) OppAvg OppN
1 Deep_Rybka_4_1.exp (d.c.s. v2) : 32 25 197.0 362 54.4 113 168 81 46.4 0 1
2 Deep Rybka 4.1 : 0 ---- 165.0 362 45.6 81 168 113 46.4 32 1
White advantage = 55.10 +/- 13.12
Draw rate (equal opponents) = 50.00 % +/- 0.00
Second round : héhé, 120 less !