It's NNUE era (sharing my thoughts)

[tomitank]

To avoid misunderstandings:
I condemn one-on-one copying with zero added value.
No need for 5 same engine.

[AndrewGrant]

To avoid misunderstandings:
I condemn one-on-one copying with zero added value.
No need for 5 same engine.

Indeed. At a minimum, train your own network. You'll pick your own unique ways of curating the data, ways of filtering it, ways of learning on it, when to do regression steps, etc. Its a tough world when asking for more than a straight copy paste job with the direction to download your networks from a Stockfish website is considered "wrong" via "facts".

Dann, you can do better. I assume you just don't know what NNUE actually is, since I've not disagreed with your posts in the past before. I even agree fully with your comment signature. What is happening here is not learning, nor is it teaching. The smart kid (SF) is letting you copy his homework. And now that you've turned your homework in, you still don't know the content.

[Dann Corbit]

I agree with Andrew. Evaluation is the soul of the engine. LMR, Null move, etc is nothing without eval.

The real power of the modern chess engine is the branching factor, not the eval.

The eval does add color and strength. But the branching factor of less than one and a half is why chess programs are running at absurd levels.

BTW, the engines have no soul. The soul is in the programmer.

Soul, from the Hebrew Nephesh means breathing thing. Chess engines don't have breathe.

I see the work that programmers put into their evaluation functions. And the code they write does belong to them.
But a sterile, mathematical NNUE answer is fine with me.
I am sorry for the authors who slaved that you don't have to slave for a NNUE eval.
But that does not change anything.

[AndrewGrant]

I agree with Andrew. Evaluation is the soul of the engine. LMR, Null move, etc is nothing without eval.

The real power of the modern chess engine is the branching factor, not the eval.

If that is the case, why are people taking SF's NNUE? The power is not in the eval you say, yet people gain hundreds to a thousand elo by copying it. Why don't we all just copy the Stockfish search tables? I'm so sick of "engine enthusiasts" that know nothing of what they speak.

I debate with myself what I should have done with Ethereal. Whether I should have sold copies. Or whether I should sell nets in the future. And my reoccurring thought is that Ethereal is at the top end, but it will never break the Leela / SF market hold. The only people using Ethereal are people that are quite invested in computer chess, and thus the only buyers would be the kinds of people on talkchess. And those people are the people who contribute ideas and insights that make things like Ethereal possible. But that is not true anymore.

But then I think of the sheer stupidity, ignorance, and otherwise degenerate views that so many on these forums hold. I don't think I view anyone on this forum as my friend anymore. I don't contribute my ideas to the programming section anymore. I've found the dozen or so people worth something on this forum, and collected them into a nice little discord. Maybe its time the Shut It Down movement goes all the way and burns this cesspool to the ground like it deserves.

[Dann Corbit]

To avoid misunderstandings:
I condemn one-on-one copying with zero added value.
No need for 5 same engine.

Indeed. At a minimum, train your own network. You'll pick your own unique ways of curating the data, ways of filtering it, ways of learning on it, when to do regression steps, etc. Its a tough world when asking for more than a straight copy paste job with the direction to download your networks from a Stockfish website is considered "wrong" via "facts".

Dann, you can do better. I assume you just don't know what NNUE actually is, since I've not disagreed with your posts in the past before. I even agree fully with your comment signature. What is happening here is not learning, nor is it teaching. The smart kid (SF) is letting you copy his homework. And now that you've turned your homework in, you still don't know the content.

There is very little learning or teaching going on, granted.
That is because it is just a big black box that spits out correct answers.
Those naughty little neurons don't even tell us what they are doing or why they are doing it.
But that is OK.
We can look at the big picture of the neural net as a tool.
I have a model of how something leans patterns.
I give it patterns and it learns them.
I use it as my little savant like Gauss used his savant.
Gauss could never do the same calculations as his savant did, even though he was truly great at calculations.
But his savant was useful to him.

There is nothing wrong with slaving away to write a colorful evaluation. Look at how fond Frank is of John Stanback's program.
I am very fond of sting,
I also like your program and many others (some very simple ones are among my favorites -- I am wild about Olithink, as everyone konws)

I am only pointing out that people who do not want to slave on an evaluation function because it is not important to them are not doing something wrong because they tape the savant into their program.

I imagine that they will be banned and scorned but I neither ban them nor scorn them in my heart because in my view they have done nothing wrong in any sort of way.

[Dann Corbit]

I agree with Andrew. Evaluation is the soul of the engine. LMR, Null move, etc is nothing without eval.

The real power of the modern chess engine is the branching factor, not the eval.

If that is the case, why are people taking SF's NNUE? The power is not in the eval you say, yet people gain hundreds to a thousand elo by copying it. Why don't we all just copy the Stockfish search tables? I'm so sick of "engine enthusiasts" that know nothing of what they speak.

I debate with myself what I should have done with Ethereal. Whether I should have sold copies. Or whether I should sell nets in the future. And my reoccurring thought is that Ethereal is at the top end, but it will never break the Leela / SF market hold. The only people using Ethereal are people that are quite invested in computer chess, and thus the only buyers would be the kinds of people on talkchess. And those people are the people who contribute ideas and insights that make things like Ethereal possible. But that is not true anymore.

But then I think of the sheer stupidity, ignorance, and otherwise degenerate views that so many on these forums hold. I don't think I view anyone on this forum as my friend anymore. I don't contribute my ideas to the programming section anymore. I've found the dozen or so people worth something on this forum, and collected them into a nice little discord. Maybe its time the Shut It Down movement goes all the way and burns this cesspool to the ground like it deserves.

NNUE added 100 Elo, not thousands.
The eval of Olithink would fit on a single page of paper.
Yet with a single thread it is within a thousand Elo of Stockfish

[AndrewGrant]

I agree with Andrew. Evaluation is the soul of the engine. LMR, Null move, etc is nothing without eval.

The real power of the modern chess engine is the branching factor, not the eval.

If that is the case, why are people taking SF's NNUE? The power is not in the eval you say, yet people gain hundreds to a thousand elo by copying it. Why don't we all just copy the Stockfish search tables? I'm so sick of "engine enthusiasts" that know nothing of what they speak.

I debate with myself what I should have done with Ethereal. Whether I should have sold copies. Or whether I should sell nets in the future. And my reoccurring thought is that Ethereal is at the top end, but it will never break the Leela / SF market hold. The only people using Ethereal are people that are quite invested in computer chess, and thus the only buyers would be the kinds of people on talkchess. And those people are the people who contribute ideas and insights that make things like Ethereal possible. But that is not true anymore.

But then I think of the sheer stupidity, ignorance, and otherwise degenerate views that so many on these forums hold. I don't think I view anyone on this forum as my friend anymore. I don't contribute my ideas to the programming section anymore. I've found the dozen or so people worth something on this forum, and collected them into a nice little discord. Maybe its time the Shut It Down movement goes all the way and burns this cesspool to the ground like it deserves.

NNUE added 100 Elo, not thousands.
The eval of Olithink would fit on a single page of paper.
Yet with a single thread it is within a thousand Elo of Stockfish

Wrong! Facts! Simple!

+180 to Ethereal and Komodo
+250 to Igel and Ruby
+300 to Andscas
+1000 to BBC

[Dann Corbit]

The graph on this page shows the effect of NNUE:
https://www.sp-cc.de/

[Dann Corbit]

I agree with Andrew. Evaluation is the soul of the engine. LMR, Null move, etc is nothing without eval.

The real power of the modern chess engine is the branching factor, not the eval.

If that is the case, why are people taking SF's NNUE? The power is not in the eval you say, yet people gain hundreds to a thousand elo by copying it. Why don't we all just copy the Stockfish search tables? I'm so sick of "engine enthusiasts" that know nothing of what they speak.

I debate with myself what I should have done with Ethereal. Whether I should have sold copies. Or whether I should sell nets in the future. And my reoccurring thought is that Ethereal is at the top end, but it will never break the Leela / SF market hold. The only people using Ethereal are people that are quite invested in computer chess, and thus the only buyers would be the kinds of people on talkchess. And those people are the people who contribute ideas and insights that make things like Ethereal possible. But that is not true anymore.

But then I think of the sheer stupidity, ignorance, and otherwise degenerate views that so many on these forums hold. I don't think I view anyone on this forum as my friend anymore. I don't contribute my ideas to the programming section anymore. I've found the dozen or so people worth something on this forum, and collected them into a nice little discord. Maybe its time the Shut It Down movement goes all the way and burns this cesspool to the ground like it deserves.

NNUE added 100 Elo, not thousands.
The eval of Olithink would fit on a single page of paper.
Yet with a single thread it is within a thousand Elo of Stockfish

Wrong! Facts! Simple!

+180 to Ethereal and Komodo
+250 to Igel and Ruby
+300 to Andscas
+1000 to BBC

Sure, they were a long way from perfection.
Because the NNUE evaluation is so wonderful, it will boost a weak eval by a mile.
That is good thing and not a bad one.
Yes, you had to labor for your eval.
The SF team probably has 10,000 hours of effort into crafting the evaluation function. That is why it was so very good.
The NNUE algorithm can replicate those 10,000 hours in an afternoon.

[AndrewGrant]

I agree with Andrew. Evaluation is the soul of the engine. LMR, Null move, etc is nothing without eval.

The real power of the modern chess engine is the branching factor, not the eval.

If that is the case, why are people taking SF's NNUE? The power is not in the eval you say, yet people gain hundreds to a thousand elo by copying it. Why don't we all just copy the Stockfish search tables? I'm so sick of "engine enthusiasts" that know nothing of what they speak.

I debate with myself what I should have done with Ethereal. Whether I should have sold copies. Or whether I should sell nets in the future. And my reoccurring thought is that Ethereal is at the top end, but it will never break the Leela / SF market hold. The only people using Ethereal are people that are quite invested in computer chess, and thus the only buyers would be the kinds of people on talkchess. And those people are the people who contribute ideas and insights that make things like Ethereal possible. But that is not true anymore.

But then I think of the sheer stupidity, ignorance, and otherwise degenerate views that so many on these forums hold. I don't think I view anyone on this forum as my friend anymore. I don't contribute my ideas to the programming section anymore. I've found the dozen or so people worth something on this forum, and collected them into a nice little discord. Maybe its time the Shut It Down movement goes all the way and burns this cesspool to the ground like it deserves.

NNUE added 100 Elo, not thousands.
The eval of Olithink would fit on a single page of paper.
Yet with a single thread it is within a thousand Elo of Stockfish

Wrong! Facts! Simple!

+180 to Ethereal and Komodo
+250 to Igel and Ruby
+300 to Andscas
+1000 to BBC

Sure, they were a long way from perfection.
Because the NNUE evaluation is so wonderful, it will boost a weak eval by a mile.
That is good thing and not a bad one.
Yes, you had to labor for your eval.
The SF team probably has 10,000 hours of effort into crafting the evaluation function. That is why it was so very good.
The NNUE algorithm can replicate those 10,000 hours in an afternoon.

You have to labour for the NNUE eval too. Unless you are a coward and reuse the code and weights that Stockfish already gave you. NNUE is not a magic bullet -- Stockfish NNUEs are a magic bullet. Everyone who has tried to replicate NNUE on their own will know this.

I will release Ethereal 13.00, and it will just by SF code verbatim. I cannot wait to be accepted with open arms, placed at the #1 spot on the rating lists, and begin showering in the profits.