A Crossroad in Computer Chess; Or Desperate Flailing for Relevance

[jp]

I'm working again now -- for chess.com actually. Its a juggling act, one which I'm losing, since I can't do 60 hours of Ethereal and 40 hours of work and 20 hours of family and 40 hours of sleep. Something has to go, and I'm pretty keen on all 4.

Maybe you should try to recruit some more people to the Ethereal team, so your 60 hours on Ethereal can come down to 30 or 40 hours.

[AndrewGrant]

I'm working again now -- for chess.com actually. Its a juggling act, one which I'm losing, since I can't do 60 hours of Ethereal and 40 hours of work and 20 hours of family and 40 hours of sleep. Something has to go, and I'm pretty keen on all 4.

Maybe you should try to recruit some more people to the Ethereal team, so your 60 hours on Ethereal can come down to 30 or 40 hours.

Easier said than done, I'm afraid. Perhaps with SF going NNUE, and the then decreased interest in the hand crafted evaluation, others might consider Ethereal. I believe the reason Alayan joined me was because adding to SF was too hard, since theres a thousand things going on and interpretations. At a time, Ethereal was a sort of blank slate for one to play with. Thats not quite the case anymore, the program is far more mature now than it was two years ago, but there is lots of room in the eval, and I'm also looking to train two dozen networks for various things.

[JohnWoe]

This is wrong, history has shown that people clone engines nearly of all levels.

Is cloning engines only a problem in computer chess or do other computer board game communities such as computer go or computer shogi have similar cloning issues?

In JavaScript selector engine competitions under SlickSpeed. I was "competing" in circa 2006-2008. There was not much drama.
That was a fierce competition tho. jQuery(dojo/prototype/mootools ...) keeping the world's #1 spot. John Resig had another pure selector engine. But my engine was faster.

I improved the code for weeks I kept finding tricks I found more and more speed. Until my JavaScript engine TonTon.js took the the World's #1 spot on SlickSpeed test. I'm looking at the source code from year 2008. The code is 4,000 lines. Pretty awful. My plan has been to clean up that code and publish it.
That competition is dead now as Chrome/Firefox give selector engine natively.

[corres]

...
I could be all wrong. I could be out of touch. But if I'm not, then the future of computer chess, the future of unique and diverse engines, depends upon all of us, as individuals, to encourage and promote new ideas while discouraging those who take from Stockfish without trying their hand at the problem. I'm already concerned when I see engines with Stockfish nets being placed onto rating lists.

In every research work there is the possibility of the partial or the total setback, I think.
Without any doubt the method of Alpha Zero Team is nearly automatized the development of chess engines.
This is the natural course of every technical innovations, earlier or later it comes a new trick what makes being superfluous the old ones.

[Rebel]

...
I could be all wrong. I could be out of touch. But if I'm not, then the future of computer chess, the future of unique and diverse engines, depends upon all of us, as individuals, to encourage and promote new ideas while discouraging those who take from Stockfish without trying their hand at the problem. I'm already concerned when I see engines with Stockfish nets being placed onto rating lists.

In every research work there is the possibility of the partial or the total setback, I think.
Without any doubt the method of Alpha Zero Team is nearly automatized the development of chess engines.
This is the natural course of every technical innovations, earlier or later it comes a new trick what makes being superfluous the old ones.

His (rightly) concern is that engine authors download the latest SF NNUE net and gain a free xx elo. Example, suppose I add NNUE (I won't, but suppose) to ProDeo and get and 150 elo gain on the rating lists because of that. 6-7 months later -- having changed not one bit on my engine -- I download the latest SF net, it gives me a free 30-50 elo, I release it and rub my hands seeing ProDeo climb and climb in the rating lists. I would say that's a very unhealthy situation and surely this (kind of things) is what's going to happen in the near future.

So while NNUE is a fantastic new development regarding increasing strength it has an unwished negative side effect and I can understand that some engine authors already now consider to stop.

[AdminX]

...
I could be all wrong. I could be out of touch. But if I'm not, then the future of computer chess, the future of unique and diverse engines, depends upon all of us, as individuals, to encourage and promote new ideas while discouraging those who take from Stockfish without trying their hand at the problem. I'm already concerned when I see engines with Stockfish nets being placed onto rating lists.

In every research work there is the possibility of the partial or the total setback, I think.
Without any doubt the method of Alpha Zero Team is nearly automatized the development of chess engines.
This is the natural course of every technical innovations, earlier or later it comes a new trick what makes being superfluous the old ones.

His (rightly) concern is that engine authors download the latest SF NNUE net and gain a free xx elo. Example, suppose I add NNUE (I won't, but suppose) to ProDeo and get and 150 elo gain on the rating lists because of that. 6-7 months later -- having changed not one bit on my engine -- I download the latest SF net, it gives me a free 30-50 elo, I release it and rub my hands seeing ProDeo climb and climb in the rating lists. I would say that's a very unhealthy situation and surely this (kind of things) is what's going to happen in the near future.

So while NNUE is a fantastic new development regarding increasing strength it has an unwished negative side effect and I can understand that some engine authors already now consider to stop.

Yes, Ed. That sums it up and I am sure both you and Andrew are correct.

[smatovic]

...
I could be all wrong. I could be out of touch. But if I'm not, then the future of computer chess, the future of unique and diverse engines, depends upon all of us, as individuals, to encourage and promote new ideas while discouraging those who take from Stockfish without trying their hand at the problem. I'm already concerned when I see engines with Stockfish nets being placed onto rating lists.

In every research work there is the possibility of the partial or the total setback, I think.
Without any doubt the method of Alpha Zero Team is nearly automatized the development of chess engines.
This is the natural course of every technical innovations, earlier or later it comes a new trick what makes being superfluous the old ones.

His (rightly) concern is that engine authors download the latest SF NNUE net and gain a free xx elo. Example, suppose I add NNUE (I won't, but suppose) to ProDeo and get and 150 elo gain on the rating lists because of that. 6-7 months later -- having changed not one bit on my engine -- I download the latest SF net, it gives me a free 30-50 elo, I release it and rub my hands seeing ProDeo climb and climb in the rating lists. I would say that's a very unhealthy situation and surely this is what's going to happen in the near future.

So while NNUE is a fantastic new development regarding increasing strength it has an unwished negative side effect and I can understand that some engine authors already now consider to stop.

The kind of competition has changed, already since appearance of Lc0 with various nets, NNUE just brought it to the broader AB CPU engines realm.

NNs replacing eval is just one part of the game, MCTS-PUCT already replaces move selection with NNs, time will come for NNs in pruning/reductions/extension and so on.

Go with the flow I say, there will be room for all kind of techniques, old and new, original or copied...

--
Srdja

[Gerd Isenberg]

...
I could be all wrong. I could be out of touch. But if I'm not, then the future of computer chess, the future of unique and diverse engines, depends upon all of us, as individuals, to encourage and promote new ideas while discouraging those who take from Stockfish without trying their hand at the problem. I'm already concerned when I see engines with Stockfish nets being placed onto rating lists.

In every research work there is the possibility of the partial or the total setback, I think.
Without any doubt the method of Alpha Zero Team is nearly automatized the development of chess engines.
This is the natural course of every technical innovations, earlier or later it comes a new trick what makes being superfluous the old ones.

His (rightly) concern is that engine authors download the latest SF NNUE net and gain a free xx elo. Example, suppose I add NNUE (I won't, but suppose) to ProDeo and get and 150 elo gain on the rating lists because of that. 6-7 months later -- having changed not one bit on my engine -- I download the latest SF net, it gives me a free 30-50 elo, I release it and rub my hands seeing ProDeo climb and climb in the rating lists. I would say that's a very unhealthy situation and surely this (kind of things) is what's going to happen in the near future.

So while NNUE is a fantastic new development regarding increasing strength it has an unwished negative side effect and I can understand that some engine authors already now consider to stop.

Looks to me there is a similar development in computer chess as in computer Shogi, where one combines for instace YaneuraOu (GPL) search (16 entries in 2019), adopted from Stockfish's search, with various evaluations, in 2019 many NNUE networks, trained by different programs, plus some sophisticated super-trained opening book. Take a look at the wcsc29 site 参加チーム (Participating teams) page the 使用ライブラリ (Library used) column ...

[corres]

...
In every research work there is the possibility of the partial or the total setback, I think.
Without any doubt the method of Alpha Zero Team is nearly automatized the development of chess engines.
This is the natural course of every technical innovations, earlier or later it comes a new trick what makes being superfluous the old ones.

His (rightly) concern is that engine authors download the latest SF NNUE net and gain a free xx elo. Example, suppose I add NNUE (I won't, but suppose) to ProDeo and get and 150 elo gain on the rating lists because of that. 6-7 months later -- having changed not one bit on my engine -- I download the latest SF net, it gives me a free 30-50 elo, I release it and rub my hands seeing ProDeo climb and climb in the rating lists. I would say that's a very unhealthy situation and surely this (kind of things) is what's going to happen in the near future.

So while NNUE is a fantastic new development regarding increasing strength it has an unwished negative side effect and I can understand that some engine authors already now consider to stop.

If a chess engine developer built the net of the Stockfish+NNE in his engine, his engine will use the knowledge of Stockfish and it will play near the Stockfish (at least in its playing stile).

[Rebel]

...
I could be all wrong. I could be out of touch. But if I'm not, then the future of computer chess, the future of unique and diverse engines, depends upon all of us, as individuals, to encourage and promote new ideas while discouraging those who take from Stockfish without trying their hand at the problem. I'm already concerned when I see engines with Stockfish nets being placed onto rating lists.

In every research work there is the possibility of the partial or the total setback, I think.
Without any doubt the method of Alpha Zero Team is nearly automatized the development of chess engines.
This is the natural course of every technical innovations, earlier or later it comes a new trick what makes being superfluous the old ones.

His (rightly) concern is that engine authors download the latest SF NNUE net and gain a free xx elo. Example, suppose I add NNUE (I won't, but suppose) to ProDeo and get and 150 elo gain on the rating lists because of that. 6-7 months later -- having changed not one bit on my engine -- I download the latest SF net, it gives me a free 30-50 elo, I release it and rub my hands seeing ProDeo climb and climb in the rating lists. I would say that's a very unhealthy situation and surely this is what's going to happen in the near future.

So while NNUE is a fantastic new development regarding increasing strength it has an unwished negative side effect and I can understand that some engine authors already now consider to stop.

The kind of competition has changed, already since appearance of Lc0 with various nets, NNUE just brought it to the broader AB CPU engines realm.

Yep, and as Andy pointed out the number of NN's are only a few, but look at Leelentein-14 vs 3 randomly chosen Lc0 nets (NN-64194, T40-1541, T60-4300) at http://rebel13.nl/dump/nn.html

NNs replacing eval is just one part of the game, MCTS-PUCT already replaces move selection with NNs, time will come for NNs in pruning/reductions/extension and so on.

That makes Andy's point even stronger.

Go with the flow I say, there will be room for all kind of techniques, old and new, original or copied...

The flow is interesting to watch, not to participate any longer.

--
Srdja