Fat Fritz 2

[Scipion]

May I ...? From an interested spectator, an ego's war rages here ! ..; with all respect due to everyone of course.

[MikeB]

The good news is that that they are acknowledge they used Stockfish upfront, so there is no surprise revelation I got this source from this server and I'm not supposed to have it, but now I have it and I can see it is clearly using SF code , yada yada yda, so we dont to have put with all that BS. It an acknowledged SF clone and their added value is the net. Will I pay $99 for a net , no, not me, I have no interest. Would I pay $25 for a far superior net, maybe.

They can't say than they are stronger than SF, they are SF. They can say their net is the best net etc., even if is not true , that is called "puffing" and it legal to boast claims that are not true unless unless the exaggeration exceeds the reality. Meaning you can say "best" , as long as you are close to being best or strongest. So in my opinion there is nothing wrong with their advertising, they did make full disclosure they are a clone of Stockfish and their rating should count as Stockfish, not as separate entity, They did release the SF code for others to use if they so desire.

To me, it looks like they made an attempt to release it in a perfectly legal way, others may disagree, but that that why we have courts and the courts can decide. I actually commend all of their disclosures, it is far better than we what we got with Houdini, and they are at least making an attempt to do it in the right way.

But the rating is a Stockfish rating, any other characterization of their rating as a non Stockfish rating is a farce. It is more of a carbon copy clone ( virtually an exact clone of SF ) than any other SF clone I am aware of. The difference is the net. I wish then well, hopefully what they have done spurs others to make even better nets. That is where the future is now, it is the creation of nets.

[marsell]

it is far better than we what we got with Houdini,

Houdini was at least really the number one for a certain period of time.

[Ozymandias]

They can say their net is the best net etc., even if is not true , that is called "puffing" and it legal to boast claims that are not true unless unless the exaggeration exceeds the reality. Meaning you can say "best" , as long as you are close to being best or strongest. So in my opinion there is nothing wrong with their advertising

It's legal and it's wrong. They say in the description that it's "arguably the strongest entity that has ever played chess", and I'm sure they could find a particular set of conditions under which that statement would be true, but their headline isn't nearly as vague. If you know, and they must, that FF2 can't beat SF under normal testing conditions, across a variety of TCs and HW configurations, you can't truthfully say it has replaced it at the top, can you?

[MikeB]

it is far better than we what we got with Houdini,

Houdini was at least really the number one for a certain period of time.

Haha - Well Stockfish-Houdini was at top

[Modern Times]

They can't say than they are stronger than SF, they are SF. They can say their net is the best net etc., even if is not true , that is called "puffing" and it legal to boast claims that are not true unless unless the exaggeration exceeds the reality. Meaning you can say "best" , as long as you are close to being best or strongest. So in my opinion there is nothing wrong with their advertising, they did make full disclosure they are a clone of Stockfish and their rating should count as Stockfish, not as separate entity, They did release the SF code for others to use if they so desire.

I just ran a match using Silver Suite Large that I found on my machine from 2011, 793 positions, at CCRL blitz time control. Similar to the one chessbase quote in their article presumably. The score was:

Code: Select all

Score of Fat Fritz 2 vs Stockfish 2021-01-11 : 106 - 99 - 1381 [0.502]
Elo difference: 1.5 +/- 6.1, LOS: 68.8 %, DrawRatio: 87.1 %

1586 of 1586 games finished.

So I think depending on test conditions their claim can be right. All that effort though for a tiny improvement if any.

But as I've said before, ratings lists aren't the use-case for engines for a normal user. Analysis is what people use their engine for. Maybe this is where FF2 shines. I'd like to see how its solving skills in test suites stack up.

[Collingwood]

Analysis is what people use their engine for. Maybe this is where FF2 shines.

Or maybe analysis is where FF2 stinks. We don't know.

[Werewolf]

They can say their net is the best net etc., even if is not true , that is called "puffing" and it legal to boast claims that are not true unless unless the exaggeration exceeds the reality. Meaning you can say "best" , as long as you are close to being best or strongest. So in my opinion there is nothing wrong with their advertising

It's legal and it's wrong. They say in the description that it's "arguably the strongest entity that has ever played chess", and I'm sure they could find a particular set of conditions under which that statement would be true, but their headline isn't nearly as vague. If you know, and they must, that FF2 can't beat SF under normal testing conditions, across a variety of TCs and HW configurations, you can't truthfully say it has replaced it at the top, can you?

My own tiny sample size, but every little helps:


Noomen short, 200 games.
5+5
SF Dev (same day as Fat Fritz released) vs Fat Fritz 2

+17 wins for SF Dev
-11
=172

for +10 elo to SF Dev.

[Modern Times]

Analysis is what people use their engine for. Maybe this is where FF2 shines.

Or maybe analysis is where FF2 stinks. We don't know.

We don't know, which is why I'd like to see its performance in some standard position solving suites so that we know the answer.

[Dann Corbit]

https://resources.whitesourcesoftware.c ... of%20price.