by hagman <google@[EMAIL PROTECTED]
>
Apr 6, 2008 at 02:33 PM
On 6 Apr., 21:17, Stig Holmquist <stigfjor...@[EMAIL PROTECTED]
> wrote:
> What formula should I use to calculate how many digits are likely to
> be missing after N draws?
>
> Will the formula 10x0.9^3N be correct or is there another one?
> Based on this formula there would be 0.424 digits missing after 10
> draws and exactly 0.5 after 9.5 draws. But these numbers are averages
> and must have a std.dev. How should I calculate it?Would the square
> root of these numbers be close enough?
To have exact formulas, use inclusion/exclusion:
Nine digits missing with P[9] = 10 * 0.1^N
Eight digits missing with P[8] = 10C2 * 0.2^N - 9*P[9]
Seven digits missing with P[7] = 10C3 * 0.3^N - 8*P[8] + 9C2*P[9]
etc.
However I would roughly expect 10/e or about 3 digits missing after
10 draws, which is substantially more than your 0.424.
How did you arrive at your formula?
