New J.A Compiles

[schahmatist]

It's also one of the reasons why I appreciate CCRL. They only test official releases. If we tracked all the development versions, there would be thousands. It's already bordering on absurd....

I used to be a big fan of SPCC, but now that his Top15 has a multitude of development versions I pretty much ignore it now.

+1

There is a point here, but the thing is that some of those dev versions participate in the strongest computer chess tournament TCEC, which I think gives them some credit to be valid versions. Also unfortunately some of those top engines do not do official releases for months and months. (if not years, like Rubichess)

[pohl4711]

It's also one of the reasons why I appreciate CCRL. They only test official releases. If we tracked all the development versions, there would be thousands. It's already bordering on absurd....

I used to be a big fan of SPCC, but now that his Top15 has a multitude of development versions I pretty much ignore it now.

I understand your point of view. But please mention, in the 2 big engine tournaments TCEC and CCC (on chesscom), the engines play with latest dev versions and some engines have very old official releases.
Example: Integral. Latest official release: V7 from February,14, 2025. But 76 patches (!) applied since then on github.
Does anybody really want to have Integral 7 in a live-engine-ratinglist, when a much better and newer version plays in the big tournaments?
I decided, the answer for me is NO.
And because my Ratinglist is so small (Top15 only), it is possibly for me, to do more testruns of the top engines, so I can provide results of latest dev versions of the engines.
And I always make sure, that all dev-versions, I use for testing, can be downloaded by everyone. If the new dev version was not taken from official github, I always provide the download-link of the engine binary, when I publish the results of this engine.

After testing Stockfish since more than a decade now, the difference between dev-versions and release-versions are not so important for me anymore. Stockfish provides lots and lots of dev versions until they reach a amout of progress (+100 normailzed Elo), and then this dev version is called official release. For me, not very exciting. For me, the progress of the engines is interesting. And you see the progress, when testing the dev versions. The more testruns, the better. In our times of free and open software, official releases are just one of a lots of dev versions.

If a dev version replaces an official release in my UHO-Top15 Ratinglist, the official release version can still be found in my full UHO ratinglist:
https://www.sp-cc.de/files/uho_full_list.txt

What I could do is: Making a second UHO-Top15 Ratinglist, using only official release versions. And providing both ratinglists. But with the additional gamepair-lists, this would mean 4 lists instead of 2. Wouldnt that be very confusing?

[Jim Ablett]

https://github.com/ericlangedijk/chessnix
Chessnix 1.2 by Eric Langedijk

Chessnix 1.2 64 JA

Windows/Linux & Android builds

Proton link:
https://drive.proton.me/urls/JDSQZMBF0W#E64evu9UuwP0

Smash link:
https://fromsmash.com/Chessnix12JA

Jim.

[SlonoSlonar]

Allow me to add my little comment to the general discussion of Jim's statement that he will only compile release versions of engines, even those from the top 5, so to speak.
These words evoked profound internal dissonance in me, caused by the fact that Jim has no problem compiling any engine on which its author is simply learning the basics of chess programming, and programming in general. It's a snap of the fingers. This entire thread is filled with engines that are of no interest to anyone except their authors and a small circle of testers. But, for example, he compiled PlentyChess only twice – the release 7.0.0 and the then-breakthrough 7.0.12. The breakthrough version 7.0.31 is now available, but Jim is completely indifferent to it and continues compiling engines with a clear conscience, running at 2450 CCRL speeds, as if anyone (except Frank Kisinski) would care. It's also worth noting that no one other than Jim has yet learned to compile PlentyChess SSE3/SSE4 at normal speed, including the engine's author himself; there are virtually no alternatives.
This psychological phenomenon genuinely baffles me, and I can say that quite frankly.

[Frank Quisinsky]

The enthusiasm for engines has waned somewhat.
My personal opinion!

Around the year 2000 my engine site are number 1 in FIDE statistic with around 4.000 diferent IPs daily. I reported about amateur chess engines, not about commercial products only. So, my opinion from my first sentence is very clear today.

I got mails from Niue, New Caledonia ... people like to play eng-eng with Arena and amateurs.
People asked me this and that.

Many persons like to development a chess programs.
Today it is easier than it was 25 years ago, no doubt about that.

All engines are all the time important.
vs. place 250 in the world, 99,9% of humans have no chance.
I think the same vs. TOP-500 available engines (not sure).

All of the engines can inspire!

With such a work, comes from hundreds of chess programmers such strong engines you like are today available.

Never forget your roots!!

I wish you many fun with the group of engines you like.
For eng-eng insinde the TOP-5 a bit boring, most of games are draw but of course I don't know what you're doing with the engines. The TOP-3 today are around 2.000 Elo stronger as a normaly club player. Maybe you can evaluate that. Perhaps you can also recognize every single Elo point for the reason that you need every day stuff ... new devs. Congratulations, you're a genius!

[Frank Quisinsky]

Discover the world of chess engines, and you will be captivated by this wonderful hobby. Virus: Chess Engines, perhaps a disease. Not a secret, I have it ... this Virus, since Chess Challenger 7 (a chess computer from the year 1979). I follow the complete way of developments, maybe this was my fault.

[GenoM]

Frank, you have made great job for our community. Arena still the best. We all here have that virus
Jim is making his nice job too and I think that Jim does not work for anyone here so nobody should make claims about his choice regarding which program he will compile.
I think also that his choice about compiling only official releases is RIGHT. Anyone's free testing author's compiles so Jim doesn't owe anyone anything, and we here should just be grateful to him for what he does.
Thank you, Jim!

[schahmatist]

Allow me to add my little comment to the general discussion of Jim's statement that he will only compile release versions of engines, even those from the top 5, so to speak.
These words evoked profound internal dissonance in me, caused by the fact that Jim has no problem compiling any engine on which its author is simply learning the basics of chess programming, and programming in general. It's a snap of the fingers. This entire thread is filled with engines that are of no interest to anyone except their authors and a small circle of testers. But, for example, he compiled PlentyChess only twice – the release 7.0.0 and the then-breakthrough 7.0.12. The breakthrough version 7.0.31 is now available, but Jim is completely indifferent to it and continues compiling engines with a clear conscience, running at 2450 CCRL speeds, as if anyone (except Frank Kisinski) would care. It's also worth noting that no one other than Jim has yet learned to compile PlentyChess SSE3/SSE4 at normal speed, including the engine's author himself; there are virtually no alternatives.
This psychological phenomenon genuinely baffles me, and I can say that quite frankly.

+1
Personally, trying to follow (and test) top 50-60 engines or so, and yes compilations of top engines are more important for me than just any engine. Also the fact that only chess.com or TCEC or some privileged testers have access to the newest dev versions, but common computer chess enthusiasts do not - saddened me a little bit. And miraculously for long time Jim gave us the same (or even newer) versions and it was a truly valuable gift. Once again thank you, Jim! Hopefully some reasonable compromise might be founded, but in any case it was a really good run!

[Jusuf]

I would like to understand, why Jim does this at all. Many thanks, whatever the answer. Jim is definitively doing more for the chess community than anybody.
But I am not qualified to recommend him what to do, cause I do not know the answer to the question above.
If I look from a user point of view, again I am not qualified to speak for others.
My humble opinion is: The help and value of Jim´s compilations is equal to the benefit in strength engines have after his compilation vs. the official available binary. And this simply means: The compilation of a current version of commits of e.g. Integral is around 100-125 points ahead of the official release. No way a compilation of an official release gives same advantage. No way compilation of engines in the 1000 CCRL range give the same ever.
Jim answered hundreds of requests. So if he is overloaded, that is our mistake.
My (selfish) recommendation for reduction of work is:
Compile TOP 50 engines of CCRL and/or TCEC Season League 2 and higher at least once every 6 months or if more than 25 new commits were made (sure that a lot of people in this thread will remind us all when reached). Rest at official release. No additional requests.

[schahmatist]

+1
Agreed! But with one important remark: if it's not fun, it should not be done. We cannot (and should not) expect anything, period. So if Jim ever runs out of his regular agenda, and feels like he has resources to do couple of top versions compilations, that he thinks are worthy, we would really appreciate it ! If not, all is good, someday (maybe) something will be released