Reconstruction of 1σ contours (68% confidence level) from the analysis of 1000 mock data for Model A (top panel) and Model B (bottom panel), for the unbiased case (cross sections for the analysis are the same as the cross sections used to generate mock data). For display purpose, the 2D probability distribution functions are estimated using a Gaussian Kernel (default method to define the bandwidth of gaussian_kde in scipy python library); the irregular shapes are related to statistical fluctuations. The colour code is related to the nuisance parameters used and the line style to the type of nuisance parameters (see Table 2). The “+” symbols represent true values. See text for discussion.