Why does an iteration that loses 24 Elo pass?

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

Moderators: hgm, Rebel, chrisw

Dann Corbit
Posts: 12542
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Why does an iteration that loses 24 Elo pass?

Post by Dann Corbit »

Id Candidate ID Current ID Pass Score Elo Delta Done Time
337 332 331 true +114 -165 =452 -24.28 true 2018-05-24 02:33:41.9339 -0400 EDT

It would not pass the SPRT type criteria. I am curious what the pass/fail criteria really is?
How is it decided if an iteration is a benefit?
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
nabildanial
Posts: 126
Joined: Thu Jun 05, 2014 5:29 am
Location: Malaysia

Re: Why does an iteration that loses 24 Elo pass?

Post by nabildanial »

Dann Corbit wrote: Thu May 24, 2018 10:00 am Id Candidate ID Current ID Pass Score Elo Delta Done Time
337 332 331 true +114 -165 =452 -24.28 true 2018-05-24 02:33:41.9339 -0400 EDT

It would not pass the SPRT type criteria. I am curious what the pass/fail criteria really is?
How is it decided if an iteration is a benefit?
The pass/fail criteria is everything that doesn't lose -150 Elo is passed.
Henk
Posts: 7220
Joined: Mon May 27, 2013 10:31 am

Re: Why does an iteration that loses 24 Elo pass?

Post by Henk »

They don't do hill climbing and not simulating annealing. Perhaps some compromise. Getting stuck in a local maximum (top) should be avoided.