So far all engines with CTG future are buggy!

[Sedat Canbaz]

Dear Chess Friends,


Just I'd like to inform you that
Today I ran a new test: Neon Tetra 2.1 with CTG future, e.g:
As CTG opening book is used: Kutup ctg, which has green lines..
The test was under Cutechess 1.2.0 plus under Fritz GUIs too,
Just to see influences about is it all right or not. as I expected,
Not so good news (in case of CTG books under Cutechess GUI),
But the good news is that, under Fritz GUI is worked all fine !!

A SMALL NOTE:
During the test under Fritz, the engines used native CTG settings
I mean simply that not played via Eng CTG future...

And just do not ask what are the results differences.. ! but at
Least I cam say, there is 150 difference (between both GUIs)
You know, if wrong...all games games are ending as wrong!!

To be more clear,
Unfortunately as same story as before happened,
CTG code seems to be buggy via Neon Tetra 2.1,
Actually all engines (with CTG future) have bugs!
Sure here I refer in case of CTG book matches under
Cutechess GUIs..e.g some engines are too earlier
Out of book or some engines prefer even the defeats..
Sure here I refer even if all lines are based on defeats!
Where just in case... I mean if under Fritz GUIs:
Then that engine is out of opening book etc. exc.
Till 25%-30% win. perc is not out of opening book..

Just another more example, see above picture please!
At 19th move, Tetra is preferred to Kd8 (under Cutechess)
But if the same position is under Fritz GUI, then the same
Opening position would be played in auto (at 19th) to Bc5
As reason, due to under all Fritz GUIs, sure in case of
With optimized book settings etc, then all engines prefer
Those opening moves which are via higher win. perc.values!
Where under Cutechess played at 19th to Kd8 (10% win.perc.)
Btw,for anyone interested..I can share the played games,
Where you'll see.. but not so important since everyone has
A PC and can check easily regarding what is going on...

And this is not all (if under Cutechess GUIs),
It does not matter with Width book settings such as 1 or 2,
Kutup CTG always prefer as Blacks 1..c5 actually Kutup book
Has as other green line (as Blacks) to 1..e5, but never used to
Play..sure no any idea why... in short, here I mean also that:
This is another serious bug with CTG book futures !

Ok..that's all for now,
And let's hope also that the new releases to be fixed!
I there any hero who will manage to fix this bug issue?
If not... I strongly suggest to be removed CTG future!
Thanks in advance

Greetings

[Sedat Canbaz]

Well, here are the played games:
https://pixeldrain.com/u/oozwXDnW

In this way,
You can check..even without running any CTG test!
Sure you are free to run a new test... NO PROBLEM !
Even I would be happy.. because why each time
I have to run ? I mean simply what about you ?)

Btw,
Kutup CTG (by Cevdet Sari) is available...where
Pyramid (by Angel Morano) is private..but I can
Help here.. in case of interest..I may create a
Bsed one which will be tuned to prefer Pyramid..
At least those opening lines vs Kutup CTG..

Plus do not forget to compare please Elo diff.
Played between Fritz GUI and Cutechess GUI !!
See picture ...not much Elo difference..hehe ))

Be aware too,
I am tired of checking...it is already enough..
It is your turn.. as next, please attach games
To validate..as here.. otherwise..under Cutechess:
I am not going to test anymore CTG opening books !
Thanks for understanding...

By the way, I wonder too:
Exeption Me..is there anyone here, who
Is aware about what is going on or not?)
I'm pretty sure not many...including programmers too!
Otherwise...I hardly doubt that they will release their
Engines with many buggy and each time (since 2023 year)!

Because so far, exc. Me
Non over these issues want to talk or to mention..sad!
Maybe all these bugs happen only on my computers ?)
I hardly doubt but let's hope these bugs appear on my PC ))

One thing more...as next:
Kutup CTG will be played by latest Brainlearn 30
Who knows? maybe the results will be as Fritz GUI ?

Greetings

[Sedat Canbaz]

GAMES:
https://pixeldrain.com/u/WoQxSseF

And as we see this time, under Cutechess:
The Elo difference is less, about 65 Elo !

And why less Elo difference ? simply because
Brainlearn 30 was too earlier out of book,
About after 5-6 moves...via B97 opening..
As reason, strictly followed CTG green settings,
Where Tetra 2.1 settings are different..prefers
Much deeper opening lines...!

And in same time that means
Kutup CTG includes much weaker lines too!
Sure I refer e.g if going deeper and deeper!

As a final conclusion (under Cutechess etc.),
Nothing correct and close results to Fritz GUI !
Sure there is nothing impossible..but it seems
It's required a lot, but 'right' work..sure in
Case of CTG and playing under NON-Frtz GUIs !

And again and again, as Blacks, via Kutup CTG:
1...e5 is never used to play by Brainllearn 30 too,
Where under Fritz GUI, quite frequent is played...
And that's other..but important differences too !

Hope helps...

Best,
Sedat

[Skynet]

Maybe all these bugs happen only on my computers ?)

Nope, you are not the only one, that's the (main) reason why SF-PB not using code for CTG book support.

[Sedat Canbaz]

Maybe all these bugs happen only on my computers ?)

Nope, you are not the only one, that's the (main) reason why SF-PB not using code for CTG book support.

Yeah..now I see...

Actually in case of with special and massive 'green' line tuning,
The played CTG books are expecting highly close to Frtitz GUI

But I am afraid to say that
Not many CTG books are created with the mentioned method ...

Greetings

[Skynet]

Maybe all these bugs happen only on my computers ?)

Nope, you are not the only one, that's the (main) reason why SF-PB not using code for CTG book support.

Yeah..now I see...

Actually in case of with special and massive 'green' line tuning,
The played CTG books are expecting highly close to Frtitz GUI

But I am afraid to say that
Not many CTG books are created with the mentioned method ...

Greetings

Yeah, as Khalid explained on github the code for CTG books is not perfect, only books with green lines can be used with "success", but again, it's under question.
https://github.com/khalid-a-omar/Polyfi ... -ctg-books

[Sedat Canbaz]

Maybe all these bugs happen only on my computers ?)

Nope, you are not the only one, that's the (main) reason why SF-PB not using code for CTG book support.

Yeah..now I see...

Actually in case of with special and massive 'green' line tuning,
The played CTG books are expecting highly close to Frtitz GUI

But I am afraid to say that
Not many CTG books are created with the mentioned method ...

Greetings

Yeah, as Khalid explained on github the code for CTG books is not perfect, only books with green lines can be used with "success", but again, it's under question.
https://github.com/khalid-a-omar/Polyfi ... -ctg-books

Well, actually here I have still some doubts about
All books with 'green' lines may be used with 'success'

And again and again I wish to mention about Kutup CTG:
Even via 'green line enabled' it did not prefer 1..e5
That's other, but as 'important question' should be..
So here it's perfectly clear that CTG code seems to be buggy!
Sure I refer for running engines via CTG future under Cutechess

On the other hand, recently
Tanick Ramz informed that he works over newer CTG code..
And let's hope for good news... )

Greetings

[Sedat Canbaz]

Btw, due to requests...because older links are dead..
Here is alternative link for Kutup CTG (by Cevdet Sarı)

And now available on Pixeldrain too:
https://pixeldrain.com/u/ikBprmZD

I think also that (if using Cutechess GUI),
Kutup CTG can be used as reference about
Next CTG future Engines are stable or not...
I mean if 1..e5 is playing (as Blacks)...

Sure this is not all,
There are many other issues, which are needed
To be optimized...but after all, time will tell..

Happy testings )