Fat Fritz 2

[Madeleine Birchfield]

CEGT have published their updated lists now, and have both Stockfish and Fat Fritz 2 on their "Best Versions" list. So they are treating them as separate entities as well.

http://www.cegt.net/40_40%20Rating%20Li ... liste.html

I'm not surprised about this, as CGET is another rating list that treats Fire and Houdini as separate entities from Stockfish as well.

[Ckappe]

I'm not surprised about this, as CGET is another rating list that treats Fire and Houdini as separate entities from Stockfish as well.

I think both Houdini and Fire were not sold and openly admitting they were StockFish??

This is admitted by ChessBase to be the cores Stockfish engine as I understand? The only difference being it uses another net as the "default"..

It makes little sense to keep adding different stockfish clones or multiple entries with all the available nets for Stockfish, just because someone makes bold claims about their net to be "stronger than the current default"

If they go down this route there will be hard to justify not including a bunch of other clones like Raubfish, Sugar AI, Crystal, cFish, etc. that also can claim to increase (or decrease) a few elo over master+deafualt net.

This Stockfish-clone should be named Stockfish in the rating lists, for honesty and clarity's sake.. Not doing this is a disservice to anyone who doesn't know the details of these engines when reading the lists.

It is also disrespectful to Stockfish creators to "pretend" to list-readers that this is another unique chess engine, much different from other clones or forks?

Will rating lists also include the Open Sourced version och this Stockfisjh-clone with another default net on GitHub? If not, what the motivation?

[Harvey Williamson]

I'm not surprised about this, as CGET is another rating list that treats Fire and Houdini as separate entities from Stockfish as well.

I think both Houdini and Fire were not sold and openly admitting they were StockFish??

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

[JohnWoe]

Anything that touch GPLv3 becomes GPLv3. Can't wiggle your way out.

If you sell the net separately then I think there's no problem.

[Ckappe]

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

I phrased myself poorly ("sold" should have been "sold as"). I meant to say that they were not sold as being stockfish-based code, therefore it would be hard for list-makers to label them stockfish on rating-lists until hard evidence exists. This case is different. No one claims this is anything else than Stockfish and should be labeled accordingly on rating-lists.

[nio]

Either Fritz 15 with SF8 (dutch version) or Fritz17 with Sf12 for 99€ ech. Hmm, cant decide.

[Harvey Williamson]

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

I phrased myself poorly ("sold" should have been "sold as"). I meant to say that they were not sold as being stockfish-based code, therefore it would be hard for list-makers to label them stockfish on rating-lists until hard evidence exists. This case is different. No one claims this is anything else than Stockfish and should be labeled accordingly on rating-lists.

Understood

[Alayan]

My last message got removed (though I can't see what would have violated the charter).

So for the record I'll state again the core of my post

Any claim of FF2 having chess style originality to at least offer something (as it's not stronger than Stockfish, which is the main marketing claim behind FF2) would need to be demonstrated by people interested in FF2, but that it is a reasonable assumption to think it clusters closely in style with SF-dev (especially at medium to long searches). I'm not going to pay 100$ to check FF2 in this regard. Dkappe's nets not clustering closely with SF's default doesn't invalidate the assumption, as their strength is also different.

And Fat Fritz should be marked as Leela (1) or Stockfish (2) derivatives by rating lists to best inform chess enthusiasts.

[dkappe]

Dkappe's nets not clustering closely with SF's default doesn't invalidate the assumption, as their strength is also different.

Night Nurse and Dark Horse are within 100 elo of the default SF12 net (60? 80?). Toga is much weaker. I did a short best move test with both of those nets on SF12 and Igel with similar results. What would one expect with what is essentially the same algorithm, just with different optimizations? Different results with the same eval would be a bug, no?

[connor_mcmonigle]

Dkappe's nets not clustering closely with SF's default doesn't invalidate the assumption, as their strength is also different.

What would one expect with what is essentially the same algorithm, just with different optimizations? Different results with the same eval would be a bug, no?

This comment betrays some serious ignorance about the nature of search functions in modern alpha-beta chess engines. Modern search functions are far cry from what most would consider an algorithm due to their reliance on a large number of ad hoc pruning, move ordering and reduction heuristics. These numerous heuristics are essential to achieving competitive performance and extremely relevant to an engine using an alpha-beta search's strength. While it is true that you will likely find many similar heuristics in top engines, top engines' search functions still vary significantly.

Different results with the same evaluation function are to be expected.

For a low fixed depth sim test, search function variety is less of a determining factor as history based move ordering heuristics are still "burning in" and some reduction/pruning heuristics are inactive at low depth. Some engines, such as Weiss (iirc), go so far as to perform a pure alpha-beta search up to the 6th search iteration.