True! Our graphical representation was to say that, for a given number of iterations, using the vanilla RB would decrease the variability/variance/dispersion by “that much” but looking at the first iteration under a given precision would indeed also work for the comparison. In the end, it should be the same in that the number of iterations needed to reach a given precision is proportional to the variance of the process. The variance and not the standard deviation then means the figures will indeed look better for the RB version. Thanks! This will be useful for the revision (soon to come!).

]]>