Has there ever been Stockfish clone better than dev version?

[Jouni]

I mean significantly stronger. And the answer is NO! Houdini 6 was at same level than SF dev. Obviously one individual cannot improve SF alone? I never download these clones - waste of time.

[Madeleine Birchfield]

I mean significantly stronger. And the answer is NO! Houdini 6 was at same level than SF dev. Obviously one individual cannot improve SF alone? I never download these clones - waste of time.

During all of July 2020, Stockfish NNUE (clone by nodchip) was 100+ elo stronger than Stockfish dev.

[corres]

I mean significantly stronger. And the answer is NO! Houdini 6 was at same level than SF dev. Obviously one individual cannot improve SF alone? I never download these clones - waste of time.

Basically you are right.
Stockfish is such a well optimized engine that as more the number of modification as stockfish will be weaker - this is my experience.
For my personal use I used to make my own derivative of Stockfish (NNUE and not NNUE) to yield a faster binary (not for test but for my analyze) the max. gain is about ~5 Elo.

[Uri Blass]

I mean significantly stronger. And the answer is NO! Houdini 6 was at same level than SF dev. Obviously one individual cannot improve SF alone? I never download these clones - waste of time.

Basically you are right.
Stockfish is such a well optimized engine that as more the number of modification as stockfish will be weaker - this is my experience.
For my personal use I used to make my own derivative of Stockfish (NNUE and not NNUE) to yield a faster binary (not for test but for my analyze) the max. gain is about ~5 Elo.

This is not correct.

If you do the right modifications to make stockfish stronger then stockfish is going to be significantly stronger but if somebody make the right modifications(for example adding NNUE was that type of change) then the stockfish team take them so it is not for a long time.

If people want to have something legal that is significantly stronger than stockfish for a long time then the only chance is by something that is not based on stockfish.

[corres]

I mean significantly stronger. And the answer is NO! Houdini 6 was at same level than SF dev. Obviously one individual cannot improve SF alone? I never download these clones - waste of time.

Basically you are right.
Stockfish is such a well optimized engine that as more the number of modification as stockfish will be weaker - this is my experience.
For my personal use I used to make my own derivative of Stockfish (NNUE and not NNUE) to yield a faster binary (not for test but for my analyze) the max. gain is about ~5 Elo.

This is not correct.

If you do the right modifications to make stockfish stronger then stockfish is going to be significantly stronger but if somebody make the right modifications(for example adding NNUE was that type of change) then the stockfish team take them so it is not for a long time.

If people want to have something legal that is significantly stronger than stockfish for a long time then the only chance is by something that is not based on stockfish.

What is not correct - beyond the provocation.

[Uri Blass]

I mean significantly stronger. And the answer is NO! Houdini 6 was at same level than SF dev. Obviously one individual cannot improve SF alone? I never download these clones - waste of time.

Basically you are right.
Stockfish is such a well optimized engine that as more the number of modification as stockfish will be weaker - this is my experience.
For my personal use I used to make my own derivative of Stockfish (NNUE and not NNUE) to yield a faster binary (not for test but for my analyze) the max. gain is about ~5 Elo.

This is not correct.

If you do the right modifications to make stockfish stronger then stockfish is going to be significantly stronger but if somebody make the right modifications(for example adding NNUE was that type of change) then the stockfish team take them so it is not for a long time.

If people want to have something legal that is significantly stronger than stockfish for a long time then the only chance is by something that is not based on stockfish.

What is not correct - beyond the provocation.

I meant
"as more the number of modification as stockfish will be weaker" is not correct but
I agree that it is only part of your sentence and you continued "this is my experience" so thinking about it again
I was wrong to write "This is not correct".

[corres]

I mean significantly stronger. And the answer is NO! Houdini 6 was at same level than SF dev. Obviously one individual cannot improve SF alone? I never download these clones - waste of time.

Basically you are right.
Stockfish is such a well optimized engine that as more the number of modification as stockfish will be weaker - this is my experience.
For my personal use I used to make my own derivative of Stockfish (NNUE and not NNUE) to yield a faster binary (not for test but for my analyze) the max. gain is about ~5 Elo.

This is not correct.

If you do the right modifications to make stockfish stronger then stockfish is going to be significantly stronger but if somebody make the right modifications(for example adding NNUE was that type of change) then the stockfish team take them so it is not for a long time.

If people want to have something legal that is significantly stronger than stockfish for a long time then the only chance is by something that is not based on stockfish.

What is not correct - beyond the provocation.

I meant
"as more the number of modification as stockfish will be weaker" is not correct but
I agree that it is only part of your sentence and you continued "this is my experience" so thinking about it again
I was wrong to write "This is not correct".

But now it is correct.

[AndrewGrant]

I mean significantly stronger. And the answer is NO! Houdini 6 was at same level than SF dev. Obviously one individual cannot improve SF alone? I never download these clones - waste of time.

No.

[mwyoung]

I mean significantly stronger. And the answer is NO! Houdini 6 was at same level than SF dev. Obviously one individual cannot improve SF alone? I never download these clones - waste of time.

It may be a waste of time to download SF clones. But it is funny that some are still buying it.

Code: Select all

ChessBase Shop

€99.90
Add to Cart
€83.95 without VAT (for customers outside the EU)
$99.06 (without VAT)
Houdini 6 Pro
PRODUCT
DESCRIPTION
SYSTEM

Houdini 6 continues where its predecessor left off, and adds solid 60 Elo points to this formidable engine, once again making Houdini the strongest chess program currently available on the market.

Languages:	German, English
ISBN:	978-3-86681-626-8
EAN:	9783866816268
System:	Windows 7 or higher
Delivery:	Download
Level:	Any

https://shop.chessbase.com/en/products/ ... or_version

[Uri Blass]

I do not think that it is a waste of time to download stockfish's clones.

1)playing strength is not the only reason to download engines.
people download engines for analysis and sting solve positions that stockfish does not solve.

2)There are people who download engines to test them.
If somebody test sting against stockfish at long time control then it may be interesting to see the results and in case stockfish wins maybe people can use the games to compose some test suite when stockfish is significantly stronger than sting.