Re: [Bug-gnubg] error rate and match winning chances

[mailinglists]

Zitat von Mueller Achim <address@hidden>:

> That's why I asked for the fibs ratings. I guess these numbers are more
> reliable regarding match winning chances.

I missed the nearby ...

http://www.fibs.com/ratings.html#calculators

Winning prob. = 1-(1/(10^((YOU-HIM)*SQRT(ML)/2000)+1))

or

1/(1+10^-((YOU-HIM)*SQRT(ML)/2000)) at http://www.bkgm.com/faq/Ratings.html

where YOU=your rating, HIM=his rating, ML=match length

For 1955.2 vs 1966.7 this results in 0.48483648 ~ 48.5%, a number  
which seems to be much more reliable.

Question: is it possible to implement this formula into the statistics panel?

Ciao

Achim