Mate in 3: 7 minutes for SF 17

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

User avatar
Eelco de Groot
Posts: 4661
Joined: Sun Mar 12, 2006 2:40 am
Full name:   Eelco de Groot

Re: Mate in 3: 7 minutes for SF 17

Post by Eelco de Groot »

Tablebases, = software, I only have Shredderbases available but hardly ever use those, only the online ones from Stefan Meyer-Kahlen in GUI are more helpful for more pieces.
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
Viz
Posts: 223
Joined: Tue Apr 09, 2024 6:24 am
Full name: Michael Chaly

Re: Mate in 3: 7 minutes for SF 17

Post by Viz »

Eelco de Groot wrote: Wed Oct 02, 2024 8:35 pm A Stockfish without nullmove would also not need a symmetrical evaluation anymore (was necessary for the verification search I believe, or something like that?) so for fans of that the old Glaurung style "Aggressiveness" and "Cowardice" if I remember the names correctly, could come back.Not that I am sure if at the moment the symmetry is not broken already because I have read that somewhere at several places. So that 'rule' was already broken then, without any Elo loss, apparently?
Stockfish with null move doesn't even have symmetrical evaluation, wth are you talking about?
User avatar
Eelco de Groot
Posts: 4661
Joined: Sun Mar 12, 2006 2:40 am
Full name:   Eelco de Groot

Re: Mate in 3: 7 minutes for SF 17

Post by Eelco de Groot »

Yes I realize it has changed, what I was referring too was back in the day there were two parameters, 'Aggressiveness' and 'Cowardice as UCI parameters, with which playing style could be adapted easily by anyone. But because after nulmove, same position but different color to move, you gained a little ELo I believe if you could just use the same stored evaluation, negating the sign. Asymmetry had to be removed. With the asymmetry this was not accurate enough. So UCI parameters stemming from Glaurung were scrapped. Afterwards I think things like optimism also changed symmetry? It's not important I was just wondering if for instance with two networks instead of one, one for each color you could introduce more variation in playing style and for optimum performance probably would have to go back to a single network. I have absolutely no idea how Patricia does that to be honest. But the two parameters from Tord were a very simple way too without much Elo loss to change HCE Kingsafety for one color for instance, apart from the litle bit simplicity lost at the start of nullmove you had to do a new evaluation.
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan