[math-fun] Brownian motion

There is a natural way to define Brownian motion in R^3 that starts and ends at the same point (1,0,0) and is a one-to-one arc. It's an interesting question to ask what are the probabilities that it forms various knots, like say a trefoil. --Dan