Interesting read about Fat Fritz 2

[Modern Times]

Open-Source GUI hype. If one of those was actually designed for the average player and their analysis, and came preloaded with some nice open-source engine...

This community would have to find some design talent, though. One only has to look at the TCEC GUI, which is extremely functional in a bleak and horrid sort of way.

CuteChess is what I use, for engine matches anyway.

[Collingwood]

This was a note I sent to Chessbase on 02/21/2021.
<snip>

Well done. They'll do as much as the consuming public allows them to get away with. If the consuming public holds them accountable, they'll have to change.

[Collingwood]

I predict soon ChessBase try to twist this universal condemnation of their misconduct into some smear campaign of a miserable person. But no, ASilver is fully aware of the scam and most of the misleading contents are either said or written by himself, and ChessBase is fully on board with those.

They have now tried to smear other companies to defend themselves:

Tord Romstad
@tordr
ChessBase and Albert Silver claim that Stockfish is used in "commercial apps like Play Magnus". I wrote the AI part of that app. It's 100% my own work, based on an early ancestor of Stockfish for which I was the only author and had the full copyright. Not like FF2 at all.

[glennsamuel32]

From Chess24

https://chess24.com/en/read/news/chessb ... fish-clone

[Collingwood]

Note: the reason for the odd number of games is that the suite used has 776 positions, which with reversed colours leads to 1552 games.

The games were attached to the article. It is plausible that they did exactly what they said they did. That doesn't, however, imply that all other testing conditions were fair.

Either the 776-game book note was added after I published the blog, or I simply missed it when I was writing.

In any case using your personal opening suite is comically bad test design. Any results that don't use some standard book should be ignored.

But honestly? Tons of authors report results on SALC, low-draw books, etc. This community is really bad at self-testing.

"Chessbase has changed their website to remove the claim that "Fat Fritz 2.0" is the "new #1" chess engine."

[Modern Times]

"Chessbase has changed their website to remove the claim that "Fat Fritz 2.0" is the "new #1" chess engine."

That is a wise move. The majority of tests say that it isn't #1 in engine matches, but given that the normal use-case for purchasers of the engine is for analysis purposes, it is hard to say how you would measure it in that context. Engine ratings are a good guide of course, an engine that rates >150 Elo above another engine say is likely to be a better analysis partner, but engines within 50 Elo of each other ? I don't know.

[gonzochess75]

"Chessbase has changed their website to remove the claim that "Fat Fritz 2.0" is the "new #1" chess engine."

That is a wise move. The majority of tests say that it isn't #1 in engine matches, but given that the normal use-case for purchasers of the engine is for analysis purposes, it is hard to say how you would measure it in that context. Engine ratings are a good guide of course, an engine that rates >150 Elo above another engine say is likely to be a better analysis partner, but engines within 50 Elo of each other ? I don't know.

Well, since SF 13 actually is #1 then this is tantamount to admitting that FF2 is just a regression. They are selling a regression.

[h1a8]

"Chessbase has changed their website to remove the claim that "Fat Fritz 2.0" is the "new #1" chess engine."

That is a wise move. The majority of tests say that it isn't #1 in engine matches, but given that the normal use-case for purchasers of the engine is for analysis purposes, it is hard to say how you would measure it in that context. Engine ratings are a good guide of course, an engine that rates >150 Elo above another engine say is likely to be a better analysis partner, but engines within 50 Elo of each other ? I don't know.

Well, since SF 13 actually is #1 then this is tantamount to admitting that FF2 is just a regression. They are selling a regression.

A Chessbase article does make a point about elo isn't the only reason why people should use an engine. As long as the engine is a peer to the best (not significantly weaker) and it gives DIFFERENT continuations a significant amount of time then it is useful for professionals (and enthusiasts).

If I could afford Fat Fritz 2 then I would buy it just to see what it thinks about certain positions and openings.

[carldaman]

There are some early indications that FF2 is a somewhat improved analysis partner (over regular SF). I can see that even with the weaker net that was freely distributed. CB is trying to make a similar case. Elo is not everything.

[the_real_greco]

There are some early indications that FF2 is a somewhat improved analysis partner (over regular SF). I can see that even with the weaker net that was freely distributed. CB is trying to make a similar case. Elo is not everything.

The (bleep) are you talking about? I would love to see these indications, if they exist.

Which since you haven't told us what these indications are (and all the CB articles have just been anecdotes), I'm going to go ahead and assume you're full of it.

Plus, come on- are you saying SILVER is able to train a net with improved analysis? How would that even work? He doesn't exactly have a huge personal knowledge of chess himself. How could he even tell if he's making progress?

So go ahead- shut me up. PLEASE.