Weak-type estimates for martingale maximal functions (Q2090747)

Let \((\Omega, \mathcal F, P)\) be a probability space equipped with the discrete-time filtration (\(\sigma\)-fields) \((\mathcal F_n)_{n\geq 0}\). Let \(f=(f_n)_{n\geq 0}\) be the adapted martingale with the maximal function \(f^{*}=\sup_{n\geq 0} |f_{n}|\). It is known that \(||f^{*}||_{p,\infty}\leq ||f||_p, 1\leq p<\infty \), and \(||f^{*}||\leq \frac{p}{p-1} ||f||_p, 1<p \leq \infty\), where \(||f^*||_{p,\infty}=\sup_{\lambda>0}(\lambda^p P(f^*\geq \lambda))^{1/p}\) denotes the weak-\(L^p\) quasinorm of \(f^*\) and \(||f||_p=\sup_{n\geq 0}(E|f_n|^p)^{1/p}\) is the \(L^p\) norm of a martingale \(f\) , and these inequalities are sharp. The main purpose of this paper is the study of sharp extensions of weak-type estimates for a martingale maximal function. Given \(1<p<\infty\) and a pair \((x,y)\) of nonnegative numbers satisfying \(x^p\leq y\), they identify the optimal upper bounds for \(|||\sup_n f_n|||_{p,\infty}\), for nonnegative martingales \(f=(f_n)_{n\geq 0}\) satisfying \(||f||_1=x \,\text{ and }\, ||f||^p_{p}=y\).