I'm disappointed with Stockfish dev.

[CornfedForever]

But Uri is not asking for the strongest chess engine. He is asking for the SF team to adopt practices that are certain to slow down Elo progress.

I don't think that is much of an argument these days. It opens wide the counter-argument that since Elo gain has slowed to such an extent post-NNUE, other options should be given some serious consideration.

Well, if the point is that some patches have harmed SF, then you or Uri can submit patches to fishtest that revert dubious changes.

Sounds a bit like you are tacitly agreeing with me about the ever slowing pace of progress by the current means.

I go back to my first posts on this topic, where I opined that people mostly seem to be complaining about the nets and that continually mixing new nets with new engine tweaks...could be an issue in determining if it is the endless engine tweaks or the nets that are responsible for what little Elo gain (or loss) is happening. That is all.

[syzygy]

But Uri is not asking for the strongest chess engine. He is asking for the SF team to adopt practices that are certain to slow down Elo progress.

I don't think that is much of an argument these days. It opens wide the counter-argument that since Elo gain has slowed to such an extent post-NNUE, other options should be given some serious consideration.

Well, if the point is that some patches have harmed SF, then you or Uri can submit patches to fishtest that revert dubious changes.

Well ... and what to do if with current fishtest methodology can't pass patch that gives better game play? (first that comes to mind is fortress detection for example)?

Then you can create your own fork!

[syzygy]

But Uri is not asking for the strongest chess engine. He is asking for the SF team to adopt practices that are certain to slow down Elo progress.

I don't think that is much of an argument these days. It opens wide the counter-argument that since Elo gain has slowed to such an extent post-NNUE, other options should be given some serious consideration.

Well, if the point is that some patches have harmed SF, then you or Uri can submit patches to fishtest that revert dubious changes.

Sounds a bit like you are tacitly agreeing with me about the ever slowing pace of progress by the current means.

I'm not contesting it because I haven't followed SF progress lately.

I go back to my first posts on this topic, where I opined that people mostly seem to be complaining about the nets and that continually mixing new nets with new engine tweaks...could be an issue in determining if it is the endless engine tweaks or the nets that are responsible for what little Elo gain (or loss) is happening. That is all.

Again, a new net is tested like any other patch. If you think that SF is no longer progressing because people are trying out all kinds of patches in the hope to improve it, then you seem to have very little understanding of SF development (and engineering in general).

Do you think SF would improve by itself if it were just left alone?

[CornfedForever]

Again, a new net is tested like any other patch. If you think that SF is no longer progressing because people are trying out all kinds of patches in the hope to improve it, then you seem to have very little understanding of SF development (and engineering in general).

Do you think SF would improve by itself if it were just left alone?

That's not quite what I'm saying. Don't worry about it.

[syzygy]

Again, a new net is tested like any other patch. If you think that SF is no longer progressing because people are trying out all kinds of patches in the hope to improve it, then you seem to have very little understanding of SF development (and engineering in general).

Do you think SF would improve by itself if it were just left alone?

That's not quite what I'm saying. Don't worry about it.

Then feel free to explain what is so special about "new nets" compared to other engine tweaks. It is all the same!

Since they are all just patches, if a "new net" is bad for SF development, then so is any other engine tweak. Since you are suggesting that new nets are bad, you are suggesting that improving SF is bad for SF development.

The only reason why you incorrectly believe that new nets are special is because this misguided thread exists...

[Sopel]

Nets can be kinda special and overfit towards specific kinds of positions, resulting in worse performance in certain cases. This is why fishtest uses a wide range of starting positions.

[Jouni]

Yesterday 24.2. there was a stunning patch Elo +11.99! But RT test was about -1 ELO . Slow progress nowadays.

[CornfedForever]

Again, a new net is tested like any other patch. If you think that SF is no longer progressing because people are trying out all kinds of patches in the hope to improve it, then you seem to have very little understanding of SF development (and engineering in general).

Do you think SF would improve by itself if it were just left alone?

That's not quite what I'm saying. Don't worry about it.

Then feel free to explain what is so special about "new nets" compared to other engine tweaks. It is all the same!

Since they are all just patches, if a "new net" is bad for SF development, then so is any other engine tweak. Since you are suggesting that new nets are bad, you are suggesting that improving SF is bad for SF development.

The only reason why you incorrectly believe that new nets are special is because this misguided thread exists...

And there is our differing way of looking at things (also...I never said 'new nets were 'bad''. Obviously they 'can' and often are better than older nets.)

To call a new net 'just another patch' because the engine and net work together is apparently your way of looking at it and if so I can't argue with that 'Borg' concept in the semantics. To me a 'patch' is sometihng you see tried in the development versions - like a LMR capture bonus 'change', or decreasing reduction for 'killer moves' or lowering the SEE pruning in qsearch....engine specific. While a new net is...well, a new net for it to use along with those actual changes to the engine like I just mentioned. You could just as well use those engine tweaks with an older net though to measure elo gain/loss. As is...it would seem hard to differentiate if the elo gain/loss was primarily a result of those engine changes...or the change in the net.

[Ajedrecista]

Hello:

Yesterday 24.2. there was a stunning patch Elo +11.99! But RT test was about -1 ELO . Slow progress nowadays.

If you are talking about this test, it was against SF 15, pretty much in line with other tests at the progress section of SF testing framework. I mean, that +11 was not against the latest development version, but against SF 15 from 18th April, 2022.

Regards from Spain.

Ajedrecista.

[Jouni]

No I was speaking about:

Author: Michael Chaly
Date: Fri Feb 24 20:14:44 2023 +0100
Timestamp: 1677266084

Search tuning at very long time control

This patch is a result of tuning session of approximately 100k games at 120+1.2.
Biggest changes are in extensions, stat bonus and depth reduction for nodes without a tt move.

Failed STC:
https://tests.stockfishchess.org/tests/ ... 625bcd7938
LLR: -2.94 (-2.94,2.94) <0.00,2.00>
Total: 13872 W: 3535 L: 3769 D: 6568 Elo -5.86
Ptnml(0-2): 56, 1621, 3800, 1419, 40

Close to neutral at LTC:
https://tests.stockfishchess.org/tests/ ... 625bcd7b8a
Elo: 0.80 +-1.2 (95%) LOS: 90.0%
Total: 60000 W: 16213 L: 16074 D: 27713 Elo +0.80
Ptnml(0-2): 24, 5718, 18379, 5853, 26
nElo: 1.82 +-2.8 (95%) PairsRatio: 1.02

Passed 180+1.8 VLTC:
https://tests.stockfishchess.org/tests/ ... 625bcdb33e
LLR: 2.95 (-2.94,2.94) <0.50,2.50>
Total: 15864 W: 4449 L: 4202 D: 7213 Elo +5.41
Ptnml(0-2): 1, 1301, 5083, 1544, 3

Passed 60+0.6 8 threads SMP VLTC:
https://tests.stockfishchess.org/tests/ ... 625bcdbdb3
LLR: 2.94 (-2.94,2.94) <0.50,2.50>
Total: 6288 W: 1821 L: 1604 D: 2863 Elo +11.99
Ptnml(0-2): 0, 402, 2123, 619, 0

closes https://github.com/official-stockfish/S ... /pull/4406