A proof of the Krylov-Safonov theorem without localization

The Krylov-Safonov theorem says that solutions to non-divergence uniformly elliptic equations with rough coefficients are H\"{o}lder continuous. The proof combines a basic measure estimate with delicate localization and covering arguments. Here we give a "global" proof based on convex analysis that avoids the localization and covering arguments. As an application of the technique we prove a $W^{2,\,\epsilon}$ estimate where $\epsilon$ decays with the ellipticity ratio of the coefficients at a rate that improves previous results, and is optimal in two dimensions.