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

az-willie wrote:

> Douglas Zare said on 11/26/2003 2:34 PM:
>
> >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.
> >
> >============
> >
> >
> I think that since I am using the luck rate per roll instead of the
> overall luck rate per game that it is already weighted. The number of
> rolls won't make a difference since I am using the per roll luck figure.
> Maybe -- heck I'm no statistician.

As a mathematician, I'm telling you: Don't average the luck
rate per roll between matches of different lengths if you
want to test that luck averages to 0. Multiply by the number
of moves to get the total luck. That should average to 0 in the
long run.

> The problem is becoming good enough to overcome bad luck.

No. When two decent players play, the winner is almost always
luckier. A good player requires less luck to win. A bad player
needs enough luck to kill an elephant. Since luck averages to 0,
the good player wins a majority of the matches, while the bad
luck is concentrated into a few losses.

If we play 3-point matches, and you always resign a gammon on
the first move of the match I need only about 25% mwc (match
winning chances) worth of luck to win, while you need 75% mwc.
I'll win 3/4 of the matches, but in every victory I will be lucky
and in every defeat I will be unlucky.

If you find the luck analysis confusing, just focus on enjoying
the game or playing well.

Douglas Zare