I've reduced a problem that I'm doing to showing that for all $n \geq r$ natural numbers, $$\sum_{i=o}^{n-r}\frac{(-1)^i}{(r+i+1) \ i! \ (n-r-i)!} = \frac{r!}{(n+1)!}$$
Or equivalently writing $n = N +r$ that either $$\sum_{i=o}^{N}\frac{(-1)^i}{(r+i+1) \ i! \ (N-i)!} = \frac{r!}{(N+r+1)!}$$ or $$\frac{1}{N!}\sum_{i=o}^{N}\frac{(-1)^i}{(r+i+1)} {N \choose i} = \frac{r!}{(N+r+1)!}$$
I thought about trying induction, with the $n=r$ ie $N=0$ case being very easy, but it seems like the inductive step wouldn't work with the $(r+i+1)$ term not fitting well with the binomial coefficient or factorials... Any help is appreciated!
Let's define $f$ as $$f(x) := x^r\sum_{i=0}^N (-x)^i\binom{N}{i} = x^r(1-x)^N.$$
Moreover, we define $$F(x):= \int_0^x f(y)dy = x^{r+1}\sum_{i=0}^N \frac{(-x)^i}{r+i+1}\binom{N}{i} = \int_0^xy^r(1-y)^Ndy.$$
Using $x=1$, we just need to prove that $$\int_0^1y^r(1-y)^Ndy = \frac{N!r!}{(N+r+1)!}$$
We'll do this inducting on $N$ (for all $r$ at the same time). If $N=0$, this is trivial.
Now, $$\int_0^1y^r(1-y)^{N+1}dy = \Big[\frac{y^{r+1}}{r+1}\cdot(1-y)^{N+1}\Big]_0^1+\frac{N+1}{r+1}\cdot\int_0^1y^{r+1}(1-y)^Ndy = \frac{N+1}{r+1}\cdot \frac{N!(r+1)!}{(N+r+2)!} = \frac{(N+1)!r!}{(N+r+2)!}.$$
So, induction is completed.