Chessbase and the engines that it sells

[Madeleine Birchfield]

There is something that is shared by many engines that are sold or were once sold by Chessbase: they are in one way or another a rebranding of an already existing engine, without disclosing the fact that they have used the code of another engine. This behaviour has been a problem with Chessbase for a while.

The Fritz engine is one example of Chessbase's dishonesty. Right now they are marketing an engine called Fritz 17, which many people in the wider chess community assume is directly derived from the original Fritz by Franz Morsch. However, what Chessbase fail to mention is that Fritz 17 shares virtually no code with previous versions of the Fritz engine, and is instead a simple rebranding of the Gingko engine by Frank Schneider. The same is true of Fritz 15 and 16, which is a rebranding of Rybka by Vasik Rajlich, and of Deep Fritz 14, which is a rebranding of Pandix by Gyula Horvath. None of these engines share a codebase with the original Fritz 1 to Fritz 13 by Franz Morsch. If Chessbase were honest with their customer base, they would have simply discontinued the Fritz engine after Franz Morsch retired, and directly sold Pandix as Pandix, Rybka as Rybka, and Gingko as Gingko.

Houdini, another engine sold by Chessbase, was already known to be a Robbolito clone by the computer chess community when Chessbase decided to first sell Houdini, and if I remember correctly, Robbolito was one of the stronger open source engines of its time, occupying a similar space in the world of open source engines of its time as does Leela or Stockfish today. Chessbase made no mention to the wider chess community that Houdini was a Robbolito derivative when selling Houdini, and thus Robert Houdart was able to get away with selling 2 versions of a Robbolito clone on Chessbase, Houdini 3 and 4. By 2016, Stockfish has become the strongest open source engine, so Houdart decided to replace wholescale the Robbolito codebase in Houdini with the Stockfish codebase, resulting in Houdini 5 and 6 being a rebrand of Stockfish 8 dev and Stockfish 9 dev that happened to be sold on Chessbase. Neither Chessbase nor Robert Houdart disclosed that they were selling a Robbolito or Stockfish for over 8 years.

The recent actions by Albert Silver in releasing Fat Fritz, a Leela clone, and Fat Fritz 2, a Stockfish clone, are only a continuation of these two behaviours by Chessbase and its engine authors of 1. replacing the original codebase of the engine with a copy of the codebase of a strong engine (which in many cases is open source) and rebranding it as a completely separate entity, failing to give proper credit to the original engine, which might involve some GPL violations, and 2. Misleading the public with its continued release of engines under the Fritz name when these engines are unrelated to the original Fritz engine by Franz Morsch.

[AndrewGrant]

Its time that people stand up and become the change that this community so badly needs, and that so many of us would like to see. That means not using ChessBase products. That means working to inform others so they too can avoid the grift of ChessBase. That means not support rating lists that give people like ChessBase the marketing power they so badly need in order to peddle their trash. I am no longer interested in CCRL in the slightest, until they take a real stance against Houdini and Fire, and now Fat Fritz 1 and 2. I'm close to completely abandoning CEGT, Ipman, FastGM, and all others that continue to play new games using Houdini and Fire.

I am trying to do my part to continue to see original, ethical work in computer chess. I'm maintaining OpenBench, which is letting a dozen authors work on their original projects, as well as helping dozens more who are running private instances. I answer many emails every week from up and coming engine authors, as well as well established ones. I offer machine power to help debug and test. I've offered my private NNUE training tools to a few people, in the hopes of getting people to explore different solutions. I've even offered it to Mark Lefler now, in hopes that they will leave the grifters in the dust bin of computer chess.

When I get around to Ethereal 13.00, I might be doing that as a commercial release. Yet I will provide all source code for the engine, and include the networks seperately as to not infringe on my own rights, and the rights I have given to users. I'll write about and talk about what I did in NNUE training differently. I'll provide as much evidence as is requested to show that what I've done is an original work. And I'll credit Stockfish with the few ideas that have made it all the way through to what I do privately.

Here is the reality: the computer chess community is in control of itself. Issues arise, and people deflect and say that its an issue for the courts. Wrong. This community runs on the court of public opinion, public discourse, and public debate. There is broad consensus that Houdini and Fire are stolen goods. Do something about it. Take them off your rating list, CCRL, CEGT, FastGM, Ipman. We know that nothing created by Albert Silver is worth our time, so don't test them, don't rate them, don't analyze with them. Albert is a nobody, don't put him in a position he has not earned.

So far, there are only two groups that have done what is so badly needed, and stood up and finally said enough is enough. TCEC has publically stated they believe Fire and Houdini are Stockfish clones, and have banned them as a result. CCC, as of recent, has not held events featuring the two engines, something which I've fought to ensure many times internally at chesscom.

CCC and TCEC are not perfect, and they never will be because there is no consensus on perfect. But they are the best we have, because of one simple fact: They have stood up to defend the honesty and integrity that seems to have diminished over the decades here.

EDIT: I'm going to post this as a new thread. Because I want to call people to action to take back this community; to boot out those who so quickly sell out the rest of us.

[drj4759]

I may start filtering out/banning Chessbase related chess engines at Chessowl in the next ratings list post.

[Modern Times]

The Fritz engine is one example of Chessbase's dishonesty. Right now they are marketing an engine called Fritz 17, which many people in the wider chess community assume is directly derived from the original Fritz by Franz Morsch. However, what Chessbase fail to mention is that Fritz 17 shares virtually no code with previous versions of the Fritz engine, and is instead a simple rebranding of the Gingko engine by Frank Schneider. The same is true of Fritz 15 and 16, which is a rebranding of Rybka by Vasik Rajlich, and of Deep Fritz 14, which is a rebranding of Pandix by Gyula Horvath. None of these engines share a codebase with the original Fritz 1 to Fritz 13 by Franz Morsch. If Chessbase were honest with their customer base, they would have simply discontinued the Fritz engine after Franz Morsch retired, and directly sold Pandix as Pandix, Rybka as Rybka, and Gingko as Gingko.

Happens in the world of cars all the time. Mazda branded as Ford, Nissan and Suzuiki and countless others. How about writing a letter to the boards of those multi-national companies and accuse them of deception and ask them to withdraw their products from sale ?

[drj4759]

Withdrawing their Chessbase contaminated chess engines may not be a reason to restore them from my rating list.

[kinderchocolate]

It's not going to work. While Chessbase is no good at chess engines, they are clearly the world leader in chess databases. Their Chessbase GUI is dominating the market like Facebook & Google. Every single professional chess authors are using Chessbase GUI.

You could remove those engines off from rating list, but nothing is achieve in reality. Chessbase simply doesn't care.

[Uri Blass]

There is something that is shared by many engines that are sold or were once sold by Chessbase: they are in one way or another a rebranding of an already existing engine, without disclosing the fact that they have used the code of another engine. This behaviour has been a problem with Chessbase for a while.

The Fritz engine is one example of Chessbase's dishonesty. Right now they are marketing an engine called Fritz 17, which many people in the wider chess community assume is directly derived from the original Fritz by Franz Morsch. However, what Chessbase fail to mention is that Fritz 17 shares virtually no code with previous versions of the Fritz engine, and is instead a simple rebranding of the Gingko engine by Frank Schneider. The same is true of Fritz 15 and 16, which is a rebranding of Rybka by Vasik Rajlich, and of Deep Fritz 14, which is a rebranding of Pandix by Gyula Horvath. None of these engines share a codebase with the original Fritz 1 to Fritz 13 by Franz Morsch. If Chessbase were honest with their customer base, they would have simply discontinued the Fritz engine after Franz Morsch retired, and directly sold Pandix as Pandix, Rybka as Rybka, and Gingko as Gingko.

Houdini, another engine sold by Chessbase, was already known to be a Robbolito clone by the computer chess community when Chessbase decided to first sell Houdini, and if I remember correctly, Robbolito was one of the stronger open source engines of its time, occupying a similar space in the world of open source engines of its time as does Leela or Stockfish today. Chessbase made no mention to the wider chess community that Houdini was a Robbolito derivative when selling Houdini, and thus Robert Houdart was able to get away with selling 2 versions of a Robbolito clone on Chessbase, Houdini 3 and 4. By 2016, Stockfish has become the strongest open source engine, so Houdart decided to replace wholescale the Robbolito codebase in Houdini with the Stockfish codebase, resulting in Houdini 5 and 6 being a rebrand of Stockfish 8 dev and Stockfish 9 dev that happened to be sold on Chessbase. Neither Chessbase nor Robert Houdart disclosed that they were selling a Robbolito or Stockfish for over 8 years.

The recent actions by Albert Silver in releasing Fat Fritz, a Leela clone, and Fat Fritz 2, a Stockfish clone, are only a continuation of these two behaviours by Chessbase and its engine authors of 1. replacing the original codebase of the engine with a copy of the codebase of a strong engine (which in many cases is open source) and rebranding it as a completely separate entity, failing to give proper credit to the original engine, which might involve some GPL violations, and 2. Misleading the public with its continued release of engines under the Fritz name when these engines are unrelated to the original Fritz engine by Franz Morsch.

I remember a lot before Fritz14 that pocket Fritz was based on shredder and later hiarcs and again chessbase did not call the engine with the right name and I did not like it:

https://en.wikipedia.org/wiki/Pocket_Fritz