According to quantum physics, it should be possible to create correlations between measurements at distant locations, which are impossible in the world of classical physics, without ``action at a distance''. The first experiment supposed to confirm this was done by Alain Aspect in Orsay, near Paris, in 1982. The experiment continues to be repeated and improved. Gregor Weihs (1998), then at Innsbruck, reported a difference of thirty standard deviations between his experimental outcomes and the null hypothesis of classical physics. However there are a number of problems with Weihs' statistics, connected to: time dependence, time variation, missing data. I will show how martingales can be used to take care of the first problems, and perhaps also discuss some new work on the missing data side, connected to the notion of ``distributed rejection sampling''. One might hope that these quantum phenomena could also be put to use in operating joint master's programs in stochastics, involving collaboration between groups at distant university departments. Certainly there remain many open problems to be solved.