Search

When a plain Monte Carlo estimate on n samples has variance σ 2 / n , then scrambled digital nets attain a variance that is o (1 / n) as n → ∞. For finite n and an adversarially selected integrand, the variance of a scrambled (t , m , s) -net can be at most Γ σ 2 / n for a maximal gain coefficient Γ < ∞. The most widely used digital nets and sequences are those of Sobol'. It was previously known that Γ ⩽ 2 t 3 s for any nets in base 2. For digital nets, Dick and Pillichshammer (2010) obtained the bound 2 t + s. In this paper we study digital nets in base 2 and show that Γ ⩽ 2 t + s − 1 for such nets. This bound is a simple, but apparently unnoticed, consequence of a microstructure analysis by Niederreiter and Pirsic in 2001. We obtain a sharper bound that is smaller than this for some digital nets. Our main finding is that all nonzero gain coefficients must be powers of two. A consequence of this latter fact is a simplified algorithm for computing gain coefficients of digital nets in base 2. [ABSTRACT FROM AUTHOR]