Weighted statistics

[Biff]

Hi Folks!

I run a weekly NFL pool where people pick the winners of
the games. Throughout the year some people may drop out
and others may join in. In past years, we've determined
the pool winner to be the person(s) that has the highest
winning percentage. This year I want to weight the process.

For example, if someone participated for the entire year
but has a lower winning percentage than someone else that
only participated for half the year but has a higher
winning percentage, I want to penalize the person that
participated less.

How do I go about weighting the winning percentage based
on the weekly participation?

I know this isn't exactly a math or stats forum but I
think it's relavent.

Thanks
Biff

 

[Myrna Larson]

You could simply multiply the percentage by the participation.

For a person with 90% wins, half a year = 0.9 * 0.5 = 0.45
80% wins, 3/4 a year = 0.8 * 0.75 = 0.60

Is that what you had in mind?

 

[Biff]

Hi Myrna!

Geez, I thought this would require some extensive
statistical equation!

That was too easy! Yes, that'll work just fine.

Thank you!

Biff

-----Original Message-----
You could simply multiply the percentage by the participation.

For a person with 90% wins, half a year = 0.9 * 0.5 = 0.45
80% wins, 3/4 a year = 0.8 * 0.75 = 0.60

Is that what you had in mind?

 

[Myrna Larson]

I think you need something "fancy" only if you don't want the final score to
be directly proportional to the participation. In that case you could set up a
VLOOKUP table with the participation (weeks? percentage?) on the left, and the
desired (arbitrary?) weighting in the 2nd column. Then multiply the % winnings
by the result of the VLOOKUP.

Hi Myrna!

Geez, I thought this would require some extensive
statistical equation!

That was too easy! Yes, that'll work just fine.

Thank you!

Biff