Effects of sampling noise propagating the parameter constraints (best fit and variance). The results are shown for the 16ν cosmology, resulting in a Gaussian posterior distribution. Left: ratio of the variance on the best-fit to the true variance of the parameters, compared to black line showing the prediction of Dodelson & Schneider (2013) (Eq. (11)). Middle: ratio of the variance of the estimated variance of the parameters to the true square of the parameter variance, compared to the solid line showing the prediction of Taylor et al. (2013)(Eq. (10)). Right: ratio of the mean estimated variance of the parameters to the true variance of the parameter compared to solid line showing the prediction of Percival et al. (2014)(Eq. (12)).