[Bug-gnubg] Re: Formula for standard error?

[Nis Jorgensen]

On 17 Aug 2002, I wrote:

> > aarVariance is calculated iteratively using formula (53) on
> > <http://mathworld.wolfram.com/Variance.html>
> 
> And in that calculation lies the error. This formula [in rollout.c]:
> 
>            aarVariance[ ici ][ j ] =
>                aarVariance[ ici ][ j ] * ( 1.0 - 1.0 / ( i + 1 ) ) +
>          ( i + 2 ) * rDelta * rDelta;
> 
> Should read 
> 
>            aarVariance[ ici ][ j ] =
>                aarVariance[ ici ][ j ] * ( 1.0 - 1.0 / i ) +
>          ( i + 1 ) * rDelta * rDelta;

In another post, Oystein wrote:

> Note that this is the standard error of a sample of the population,
> and not the standard error of the whole population. That's the
> difference!

Well, I was aware that this might (and should be) the case. But that 
doesn't change my belief that the calculation of it is wrong. 

I would like someone else with the necessary background to look at 
those lines in rollout.c, and tell me whether I am right.-- 
Nis Jorgensen
Work phone: +31 (20) 52 36 202
Private mobile: +31 (6) 1496 3629