Let $S=(S(n))_{n\geq 0}$ be a symmetric random walk starting at zero. Prove that for all $m \geq 1$ $$ 2\mathbb{P}(S(1)>0, \ldots, S(2m)>0)=\mathbb{P}(S(1)\geq 0, \ldots, S(2m)\geq 0).$$
There is a theorem that says that if $S(0)=0$ then $$\mathbb{P}(S(1)\neq 0, \ldots, S(2m-1)\neq 0, S(2m)=b)=\frac{|b|}{2m}\mathbb{P}(S(n)=b).$$
Using this theorem, I tried $$\mathbb{P}(S(1)>0, \ldots, S(2m)>0)=\sum_{k=1}^\infty \mathbb{P}(S(1)>0, \ldots, S(2m-1)>0, S(2m)=k) =\sum_{k=1}^\infty \frac{k}{2m}\mathbb{P}(S(2m)=k) =\sum_{k=1}^\infty \frac{k}{2m} \binom{2m}{1/2(2m+k)}(1/2)^{2m}.$$
I am stuck here. Perhaps using this theorem was a bad idea?
To prove this identity, one simple observation is enough.
image source
Let ${\cal P}_{2n}$ and ${\cal N}_{2n}$ be the sets of paths satisfying $\{S(1)>0, \ldots, S(2n)>0\}$ and $\{S(1)\geq 0, \ldots, S(2n)\geq 0\}$ respectively for all $n \in \Bbb{N}$.
$$2\Bbb{P}({\cal P}_{2m}) = 2\;\frac{|{\cal P}_{2m}|}{2^{2m}} = \frac{|{\cal N}_{2m-1}|}{2^{2m-1}} = \Bbb{P}({\cal N}_{2m-1})$$