Notice that $$F_r(z) = \frac{1}{(1-z)^{r-1}} - \sum_{k=0}^{r-1} \left(\frac{z}{1-z}\right)^k$$ and therefore for $r\geq 4$ and $n\geq 1$, we have \begin{split} [x^n]\ F_r(z) &= \binom{n+r-2}{r-2} - \sum_{k=1}^{r-1} \binom{n-1}{k-1} \\ & = \binom{n+r-2}{r-2} - \binom{n-1}{r-2} - \binom{n-1}{r-3} - \sum_{k=1}^{r-3} \binom{n-1}{n-k} \\ &\geq \binom{n+r-2}{r-2} - \binom{n}{r-2} - \sum_{k=1}^{r-3} \binom{n-1+r-3-k}{n-k} \\ &= \binom{n+r-2}{r-2} - \binom{n}{r-2} - \binom{n+r-4}{r-3} \\ &> \binom{n+r-2}{r-2} - \binom{n}{r-2} - \binom{n+r-3}{r-3} \\ &= \binom{n+r-3}{r-2} - \binom{n}{r-2} \\ &> 0. \end{split}
Max Alekseyev
- 34.3k
- 5
- 74
- 152