Page 3 of 3
Re: IID (skipped when in check and Score greater than alpha)
Posted: Tue Jul 25, 2017 9:05 pm
by tomitank
Only in a PV node?
1. is without IID?
Time control?
Re: IID (skipped when in check and Score greater than alpha)
Posted: Tue Jul 25, 2017 9:15 pm
by hgm
AlvaroBegue wrote:I don't really know what you mean by robustness. How do you define it? And how do you measure it? (To be clear, these are not rhetorical questions: I genuinely want to know your answers.)
'Robust' means that it always works, even in pathological cases.
One way to measure it would be to take the worst ratio of the node counts of two subsequent iterations, on a large set of test positions. Large ratios indicate pathological situations (e.g. where it has to change PV).
Re: IID (skipped when in check and Score greater than alpha)
Posted: Tue Jul 25, 2017 9:16 pm
by AlvaroBegue
hgm wrote:AlvaroBegue wrote:I don't really know what you mean by robustness. How do you define it? And how do you measure it? (To be clear, these are not rhetorical questions: I genuinely want to know your answers.)
'Robust' means that it always works, even in pathological cases.
One way to measure it would be to take the worst ratio of the node counts of two subsequent iterations, on a large set of test positions. Large ratios indicate pathological situations (e.g. where it has to change PV).
That's a good idea. I should make a test for that so I can run it after every change.
Thanks for the idea!