the "Importance" weight of each sample is determine by the likelihood
w=dbinom(s,n,p.prior)