I did some regressions of benchmark scores vs cubeless money play scores:

http://compgammon.blogspot.com/2012/02/bug-in-benchmark-calculation-fixed.html

Short story: almost all the money play score variation comes from Contact 
benchmark score. There's a strong dependence of money score against Crashed 
benchmark score if you do a linear regression, but when you do a multivariate 
linear regression, that dependence mostly disappears - it was there just 
because Crashed and Contact benchmark scores are highly correlated.




On Feb 20, 2012, at 7:14 AM, Ian Shaw wrote:

> Hi Philippe,
> 
> I don't think we do. Øystein performed tests with some nets he was 
> experimenting with. He only looked at cubeless play, and  I think he only 
> tested the Contact nets.
> 
> -- Ian
> 
> -----Original Message-----
> From: Philippe Michel [mailto:[email protected]] 
> Sent: 15 February 2012 20:07
> To: Ian Shaw
> Cc: [email protected]
> Subject: Benchmark error to ppg loss
> 
> On Tue, 14 Feb 2012, Ian Shaw wrote:
> 
>> 1 point of contact-benchmark error is about 156.5 micropoints/game
> 
> This is for checkers play errors only apparently. Do you have a similar ratio 
> for cube errors and for the crashed net benchmark ?
> 
> _______________________________________________
> Bug-gnubg mailing list
> [email protected]
> https://lists.gnu.org/mailman/listinfo/bug-gnubg


_______________________________________________
Bug-gnubg mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/bug-gnubg

Reply via email to