[BLML] Law 78, Neuberg and Ascherman

Steve Willner swillner at nhcc.net
Wed Apr 16 15:50:43 CEST 2014


On 2014-04-16 8:50 AM, Peter Smulders wrote:
> Nothing is said about how to convert the matchpoints to a percentage score,

No one needs to do this.

> let alone how to do this in case of boards with different tops.

Another lacuna is what to do when contestants play differing numbers of 
boards.

> Is it not time that Ascherman and Neuberg are explicitly incorporated in
> the laws?

Perhaps there should be default rules in the Laws, but RAs should have 
authority to adopt their own.  There is no mathematical basis for 
Ascherman/Neuberg (technically they are biased estimators), and I don't 
know that anyone has a proper formula.  Just as an example, present 
arrangements for handling differing tops and differing numbers of boards 
are contradictory.  I think it should be possible to derive a proper 
formula based on Stein's Paradox, but so far as I know, it hasn't been 
done.  In the meantime, Ascherman/Neuberg is a reasonable approximation.

More on this subject is at
http://www.desjardins.org/david/factor.txt

> I have put some notes on this subject on the internet, showing that
> Neuberg and Ascherman are not all that difficult to understand, see
> http://www.pjms.nl/NEUBERG/index_en.htm

I agree with you that they are not difficult to understand or apply.



More information about the Blml mailing list