SlowChess Blitz Classic 2.0

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

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
jonkr
Posts: 18
Joined: Wed Nov 13, 2019 12:36 am
Full name: Jonathan Kreuzer

Re: SlowChess Blitz Classic 2.0

Post by jonkr » Sat Mar 21, 2020 7:27 pm

Oh the _noPop was still the old compile (modified 19th), I guess it didn't get updated in the zip, I rezipped/reuploaded it so that one should hopefully work now.

Modern Times
Posts: 2476
Joined: Thu Jun 07, 2012 9:02 pm

Re: SlowChess Blitz Classic 2.0

Post by Modern Times » Mon Mar 23, 2020 1:43 pm

jonkr wrote:
Fri Mar 20, 2020 11:44 pm

- On the plus side, this should be the largest strength increase between versions so far. Scored about +100 elo in 60 + .5 self-play. (Might be less against other engines or at different time controls.)
I'm getting very close to this on the chess960 games played so far.
.

Opinions expressed here are my own, and not necessarily those of the CCRL Group.

jonkr
Posts: 18
Joined: Wed Nov 13, 2019 12:36 am
Full name: Jonathan Kreuzer

Re: SlowChess Blitz Classic 2.0

Post by jonkr » Tue Mar 24, 2020 8:41 pm

I checked the FRC list and I'm happy to see the first external testing results are consistent with what I was hoping. (Was +91 elo when I checked, and I think my results from doing self-play tests at super fast time controls will overestimate differences.)

Close to Shredder too, I remember Shredder winning a lot of world championships, would be interesting to make something that ranks above it. Maybe in a couple months I will try some more improvements, even though current top programs so far ahead.

Before release I did some FRC testing against Critter with about the same results. I also played some FRC games to add to tuning positions, which I think did lead to an improvement, but I don't play enough test games sure of anything.

Modern Times
Posts: 2476
Joined: Thu Jun 07, 2012 9:02 pm

Re: SlowChess Blitz Classic 2.0

Post by Modern Times » Tue Mar 24, 2020 9:59 pm

jonkr wrote:
Tue Mar 24, 2020 8:41 pm
I checked the FRC list and I'm happy to see the first external testing results are consistent with what I was hoping. (Was +91 elo when I checked, and I think my results from doing self-play tests at super fast time controls will overestimate differences.)
Yes, well remember we use Bayeselo and it is +91 Elo currently as you say. But if you run the list through Ordo you get +101 Elo. Take your pick :D
.

Opinions expressed here are my own, and not necessarily those of the CCRL Group.

Post Reply