>You misconstrued. Reject as little as possible (<> reject little) was the motivation with respect to more naive random walks which suffer at dimensionality.

Misunderstanding on my part; apologies.

Can you clarify the statement on dimensionality? Moment calculations converge at a sharp rate of 1/sqrt(n). Riemann sums as approximations to integrals (moments) suffer from the curse of dimensionality.