On the probability that the maximum of a polynomial is at an endpoint of an interval (Q2658369)

Consider a polynomial of degree \(n\), \[ p(x)=\sum_{k=0}^{n}a_kx^k,\] where the coefficients \(a_k\), \(k=1,\dots,n\), are independent and identically distributed random variables, on the symmetric interval \(I=[-a,a]\). The aim of this paper is to study the probability that the maximum of \(p\) is attained at the endpoint of interval \(I\). Explicit formulas for \(n=1,2,3\) are found. The formula for \(n=3\) leads to a lower bound for the true probability. It occurs, due to certain numerical examples, that similar ideas are useful for solving the problem of finding average of Markov factors \[ M(p)=\frac{\Vert p'\Vert_{[a,b]}}{\Vert p\Vert_{[a,b]}},\] where \(-\infty<a<b<\infty\), \(p\) is a polynomial with its first derivative \(p'\) and \(\Vert f\Vert_{[a,b]}=\max_{a\le x\le b}\vert f(x)\vert \).