#### Mentz114

Gold Member

- 5,307

- 242

I thought I might prove it by calculating ##\Delta_k = \sum_k sp^s - \bar{s}## and showing that this tends to 0 as ##k\rightarrow \infty##. The general term is ##\Delta_k = \frac{n^{k+1}\left((k+1)n-km\right)}{m^k(m-n)^2}## and using ##k\approx k+1## this gives ##\frac{n}{n-m}kp^k##.

I suspect this is not a proof and if not is there another way ?