A very elementary proof of a probabilistic limit relation (Q753045)

Let be \[ a_ n=e^{-n}\sum^{n-1}_{k=0}n^ k(k!)^{-1}. \] The Central Limit Theorem of the probability theory implies that \[ (*)\;\lim_{n\to \infty}a_ n=1/2 \] (\(a_ n\) is the probability that the sum on n independent, equally distributed Poisson random variables with parameter \(\lambda =1\) is smaller than the mean value). In this paper the limit relation (*) is proved by using completely elementary tools. Besides other tricks, inequalities \[ (1+1/n)^{n+1/3}<e,\;n!e^ n>n^{n+1/3},\;n=1,2,... \] are used.