SPCC: Testruns of Rebel EAS 2 and Rebel Extreme finished

yes and no, just meant to express that Stashchess and Alexander are from the same "author". Personally, I think it's a shame that a Stockfish clone is allowed to take part in the ICGA World Championship. Every test of another Stockfish clone will lead to more such clones. There will be no such tests at CEGT. (Peter M. ?)

new: https://lczero.org/blog/2024/10/why-we-withdrew-from-wccc/

Dio wrote:

 (Peter M. ?)

Martan, dear Dio

And no, I won't discuss, why i don't think there's anything wrong with a Stockfish- derivative playing at an ICGA- tournament, if no one of SF- team wants to participate and if it's ok with ICGA. Andrea Manzo calls his engine(s) SF- derivatives by his own definition and nomenclature too and gives the code along with the engines, so what?

GPL isn't a copyright primarily, it's a copyleft, meaning people aren't only allowed to use it for further development, they are invited and prompted by the license to do so. SF wouldn't be, what it is, whithout that license and the many users spending their hardware and their manpower to develop the code and the NNUE.

That Lc0 has withdrawn I'm sorry too, but that's none of my business neither, regards

Thanks Peter, I am Jörg Burwitz.

Alright, Jörg, thought so anyhow and of course you know, that I don't mean any offense with what I'm writing above (edited still while you had already answered again), not against you and not against anybody else who might be of other opinion about the case.
Anyhow ICGA is and was a touchy subject and theme of it's own since quite a time (50 years or so?), wasn't it?

This championship isn't about competition. It is the last one. It is about participating because computerchess on pc is finished like Kasparov was finished against deep blue.

peter wrote:

To see, if the avatar- setting causes trouble, I had another one match Alexander Santiago- Alexander 2.0, this time both engines using the alew.avt- setting, again 3'+1" single thread, UHO 2024 900-990cp 6mvs:

Score of SantiagoAlew vs Alex2Alew: 161 - 79 - 260 [0.582]
Elo difference: 57.5 +/- 21.1, LOS: 100.0 %, DrawRatio: 52.0 %
500 of 500 games finished.

And with EAS- tool:

Code:

                                 bad  avg.win
Rank  EAS-Score  sacs   shorts  draws  moves  Engine/player
-------------------------------------------------------------------
   1     59380  06.83%  10.56%  20.77%   74   SantiagoAlew  
   2     45200  05.06%  15.19%  24.23%   76   Alex2Alew  
-------------------------------------------------------------------
*** Average length of all won games:     75 moves

I tried 2 machines, I tried switching High Tal not to true. I tried a shorter .avt filename (alew.avt). All these 3 tries versus Alex 2.0 alew100 give a quick 0-20 or 1-19 or so.
And I tried both engines with default setting: 1-19 after 20 games.
Santiago avx2 is defintly complete broken.

And your results seem weird, too: Much higher draw-ratio when using the alew-setting (52%) (compared with default (31.4%) )???
And +161 Celo for Alex Santiago vs Alex 2 both default? Impossible, I would say, for a stockfish-clone-HCE engine. This would be the 7th wonder of the world, if this would be true. And Alex Santiago would be way stronger than Stockfish final HCE... do you really believe this?? I dont.
Mention, there are only small changes from Alex 2 to Alex Santiago (from github):
Stockfish Patch
Oct 12, 2024
Make low ply history size fixed
Various improvements and some bugs fixed.
So, this gives +161 Celo ? Come on...

So, I keep Alexander 2 for now. In cutechess-cli, Santiago plays only crap. Additionally, I doubt, that huge Celo-gains are possible here, because Alex is a HCE-Engine. And my (restarted) testrun of Alexander 2 alew100 gives around -130 Celo to Stockfish final HCE. This is completely legit and much more Celo can not be expected by a clone-engine, when using a very aggressive setting:
Komodo 14.1 (HCE too) looses around -140 Celo, when switching its playing-style to aggressive, in my old SPCC-Ratinglist.
Fits perfectly well.
So, even if Santiago would not be buggy, I would not expect much Celo-gain, compared to Alex 2.0. Definitly not +161.

pohl4711 wrote:

In cutechess-cli, Santiago plays only crap.

Maybe that's the point, I used CuteChess GUI. Second one difference: my used compile was bmi2, not avx2 like in your trials.
Anyhow here are the games:

https://drive.google.com/drive/folders/1Q6shD-KD-vodapaIDGvWQp9jwVpBUR9l?usp=sharing

peter wrote:

pohl4711 wrote:

In cutechess-cli, Santiago plays only crap.

Maybe that's the point, I used CuteChess GUI. Second one difference: my used compile was bmi2, not avx2 like in your trials.
Anyhow here are the games:

https://drive.google.com/drive/folders/1Q6shD-KD-vodapaIDGvWQp9jwVpBUR9l?usp=sharing

Alexander 2.0 alew setting testrun aborted, after I was imformed about the inacceptable behaviour of the author of Alexander on Stockfish discord ("If you want to ban me for this, this is nazism"). For me, coming from Germany, blaming anyone or anything as "nazism" is completely unacceptable.

Another low point by Shashchess author?

Admin wrote:

Another low point by Shashchess author?

Yes.
https://discord.com/channels/435943710472011776/813919952713154590/1117362334357266512

This is the link to the post of him on discord.

For me, this is completely unacceptable. Blaming anyone or anything as "nazism". The posting is from 2023, but I did not know that, until today. My bad.

Just my opinion, of course.

Awful, just awful.