Let $x_1,x_2,\cdots,x_n$ be a set of $n$ observations where $x_i\in(-a,a)\,,i=1,2,\cdots,n$ for some $a>0$. Suppose $x_{(1)}<x_{(2)}<\cdots<x_{(n)}$ are the ordered observations. Is it true that $$\max(-x_{(1)},x_{(n)})=\max_{1\le i\le n}|x_i|\quad?$$
I suspect this is true, and it might be obvious. But I could not put together a rigorous proof.
The reason I felt it is true was that if $-a<x_1,x_2,\cdots,x_n<a$, then $x_{(1)}>-a$ and $x_{(n)}<a$, together implying $a>\max(-x_{(1)},x_{(n)})$.
Again, $-a<x_1,\cdots,x_n<a\implies 0<|x_1|,\cdots,|x_n|<a\implies \max|x_i|<a$.
So we have $a>\max(-x_{(1)},x_{(n)})$ and $a>\max|x_i|$.
But I cannot conclude that both lower bounds of $a$ are equal from here.
We certainly have $$\max(-x_{(1)},x_{(n)})=\frac{1}{2}\left(-x_{(1)}+x_{(n)}+|-x_{(1)}-x_{(n)}|\right)$$
I would also like to know what happens if I include the terminal points, i.e. what if $x_i\in[-a,a]$ for all $i$ ? Will the result still hold?
As $-x_1\le |x_1|$ and $x_n\le |x_n|$, certainly $$ \max\{-x_1,x_n\}\le\max_i|x_i|.$$ There exists $k$ such that $\max_i|x_i|=|x_k|=\max\{-x_k,x_k\}$. From the given ordering, $x_1\le x_k\le x_n$ and $-x_1\ge -x_k\ge -x_n$, hence $$\max_i|x_i|=|x_k|=\max\{-x_k,x_k\}\le \max\{x_n,-x_1\}.$$