Re: this sounds like a whine I know.....

az-wille:-

Although I play on the 'Grid, I keep a Excel spreadsheet database of Snowie
analysis data for each match, and for curiosity's sake, I initially included
columns for luck rate (in normalised emg).  All the data is weighted by move
number.  After only about 100 or so matches, the luck rate stabalised close
to zero, such that there was no indication that I had been anything other
than averagely 'lucky'.  Obviously this is not rigourous, but I think a few
hundred matches would be enough to spot statistically significant trends.

Adam


"Douglas Zare" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]
>
>
> az-willie wrote:
>
> > Over a range of matches, if FIBS is truly random, everyone's luck factor
> > should equal 0 ( zero ).
>
> You need to be careful with this. Luck should average to 0
> only if you weight the matches appropriately. I believe you
> should weight your luck rate by the number of your rolls.
>
> It's like tossing a fair coin, computing the percentage of
> the tosses that are heads. If you decide to stop after the
> first head comes up, then you will get 100% half of the
> time, 50% 1/4 of the time, 33% 1/8 of the time etc. So,
> if you average by match, you would find too many heads.
> You need to weight the short matches less.
>
> Douglas Zare
>
>
>