TalkChess.com

Since most other parameters on "go" impose some search restriction, I treat the abscence of any such restrictions as "go infinite". However, you may set any default you like, such as a default search time. If the GUI wants to have a specific behaviour, it has to set the appropriate parameters - whic...

Fun fact: in one of the games, my engines reached the KBN-K endgame, so this does actually happen. Since it doesn't support EGTBs, it's good that it has some KBN-K code to win this.

is -45Elo for Minic ... How many games? Does this happened to someone else ? No, just checked at a thousand games. However, the GCC-10 generated binary is 4.5% faster for me. Any countermeasure (apart from falling back to gcc9)? Run your engine with a single worker thread. Use a few positions to ca...

It's something I don't understand: you have the source-code in front of you, you can READ what's going on right? You know how an instruction behaves, why use stepping? I also find debuggers mostly useless, except for rare cases such as a buffer overflow that overwrites something else which crashes ...

syzygy wrote: ↑

Tue Dec 29, 2020 2:20 pm

Because SF prefers 0.00 over -0.01.

With the standard contempt of 24, that would be preferring 0.00 over -0.25, no?

Ok. may be as i said telling SF the "difference" between a SURE draw by rule of repetition and an evaluated draw (other than egtb) could open a very tiny window of improvment The alternative to allowing repetition in the example positions wasn't an endgame draw. It would have been a clearly worse-t...

What do you mean by half-open window with fail high/low against no window at all Testing two versions against each other with different root window handling. One uses a +/- 50 centipawns window and re-searches with half-open window upon fail high/low. The other always searches with fully open window.

If SF always search for opponent's possible checks to its king's in say next n moves we can then assume that when a 3 repetition by forced checks occurs SF is never taken by surprise then ? Yes. Stockfish knows that this will be draw through repetition. In particular, checks are extended in search ...

Jouni wrote: ↑

Tue Dec 29, 2020 8:30 am

OS: CentOS 8

That's probably not the best choice, given that the EOL date for CentOS 8 has just been preponed from 2029 to 2021, after which CentOS will become a rolling release testbed. CentOS 7 is still supported until 2024, provided that IBM/RedHat won't axe that, too.

@ras : when you Say a draw is a draw, are you saying that when SF evaluates a position as a draw he is always right ? The problem isn't the draw. It's that the alternatives Stockfish can see would be worse, and not just by a tiny little bit. That's what the contempt is for, to prevent the engine he...