Stockfish vs ChessBase hearing a “meh” affair

[mvanthoor]

Aha! Google indeed showed his name in connection with Fritz 18, but I did not realise he is the author of Ginkgo.

Fritz 18 is just a rebranded Ginkgo, just like Fritz 15-16 were just rebranded Rybkas.

Fritz 14 was actually Pandix, by Giyula Horvath.

The last "real" Fritz by Frans Morsch was Deep Fritz 13. Nowadays, "Fritz" is just the GUI, with some third-party engine running under it.

[Ovyron]

Fritz 14 was actually Pandix, by Giyula Horvath.

Right. I'm actually glad this happened because I got to enjoy engines that would otherwise had remained private, and even a new version of Rybka (not the latest which wouldn't even turn in my CPU.)

Current wishlist:

Someone has a bright idea that is better than NNUE but keeps it private, makes own decent search and is able to overtake Stockfish, then it becomes Fritz 19. Yeah, huh, I'm kind of bored of seeing Stockfish at the top...

Nodchip lost his chance for this, Stockfish's classical eval would have made leaps and bounds to become number one and defeat NNUE's implementation of Fritz 18 (in that alternate timeline), one can only imagine where we would be at now.

[mvanthoor]

Fritz 14 was actually Pandix, by Giyula Horvath.

Right. I'm actually glad this happened because I got to enjoy engines that would otherwise had remained private, and even a new version of Rybka (not the latest which wouldn't even turn in my CPU.)

Current wishlist:

Someone has a bright idea that is better than NNUE but keeps it private, makes own decent search and is able to overtake Stockfish, then it becomes Fritz 19. Yeah, huh, I'm kind of bored of seeing Stockfish at the top...

Nodchip lost his chance for this, Stockfish's classical eval would have made leaps and bounds to become number one and defeat NNUE's implementation of Fritz 18 (in that alternate timeline), one can only imagine where we would be at now.

I haven't really cared for the latest and greatest engines at the top for some time.

I'd rather follow development of new engines in the 1500-3000 range. Especially the ones in the 1600-2400 CCRL range play chess I can actually _understand_. Sometimes Stockfish shows you a move which looks like "what?" (such as a knight backward from f6 to e8), and then you follow the line and the knight just "happens" to "luckily" cover d6, making a 28 move line work. Or worse (which already happens to me with engines in the 2000 range): you play a line, think you have a massive tactical shot somewhere around move 8 in the future, and then the engine plays some sort of in-between move no human would EVER consider which makes your beautiful, brilliant sacrifice NOT work.

I've had this so often... lines of 8, 9, 10 moves deep of which I KNOW would have been winning against any sub-2100 player (because: if I didn't see the in-between move, they probably also wouldn't), but the engine NEVER misses it.

That's the one pet peeve I have: you can't out-tactic an engine, except if it has a functionality to allow that, such as Shredder, with its user-definable ELO setting.

Therefore I often also don't analyze with Stockfish; I often still analyze with Deep Fritz 13 or one of the older "classical" engines in the 2800 ELO range, because I can at least understand what they are trying to do on the board.

It is as Larry Kaufmann and Modern Times said in another thread: current-day top engines are so strong that, if they come out of the opening in a balanced position, it will stay balanced and all games will end up in a draw; therefore computer chess is almost dead at the top, but in the lower regions, it's very much alive.