Looks like this method converges at root(N) same as a direct monte-carlo [1]

Anyone got a stochastic method that converges quicker?

[1] http://sohcahtoa.org.uk/pages/maths_montecarlo.html