(a) Triangle plot of the Bayesian posterior and (b) corresponding profile likelihood functions of the sampling parameters. The data have been fit using the MBPL approach with three free breaks (﹩{N}_{{\rm{b}}}=3﹩). The posterior median is depicted by the solid green line. Maximum likelihood parameter values are depicted by the dashed blue lines. Dot-dashed black lines show the uncertainty estimates. The scalings of the x-axes have been chosen in accordance with the priors (flat or log-flat). The parameter AS is given in units of ﹩{\rm{s}}\;{\mathrm{cm}}^{2}\;{\mathrm{sr}}^{-1},﹩ while the units of the breaks ﹩{S}_{\mathrm{bj}}﹩ and the integral flux Fiso are ﹩{\mathrm{cm}}^{-2}\;{{\rm{s}}}^{-1}.﹩ (a) The diagonal shows the marginalized posteriors for single parameters, while two-dimensional correlations between parameters become evident from the off-diagonal plots. Uncertainties have been calculated from the 15.85% and 84.15% quantiles. Note that contours have been chosen for visibility puposes only and do not represent a specific confidence interval. (b) The profile likelihood has been derived from the posterior samples. Black circles depict bin centers. Uncertainty estimates are 68.3% CL. In case the profile likelihood was not sufficiently constraining, uncertainty estimates have been approximated with the limits of the sample data.