The Bayesian Subset Simulation (BSS) algorithm is a recently proposed approach, based on Sequential Monte Carlo simulation and Gaussian process modeling, for the estimation of the probability that $f(X)$ exceeds some thresold $u$ when $f$ is expensive to evaluate and $P(f(X)>u)$ is small. We discuss in this talk the bias an variance of the BSS algorithm, and propose a variant where the bias-variance trade-off is automatically tuned.