Difference between ProfileLikelihood and MCMC Calculator


We are trying to perform and un-binned fit in the invariant mass distribution for the Z’ analysis using four nuisances, we are comparing the results obtained using a Voigtian with the ones obtained using a Crystal-ball convoluted with a BW using RooFFTConvPdf, we have observed that the Profile likelihood calculator throws more or less the same result for both the Voigtian and the Gaussian, nonetheless when comparing the output of the MCMCCalculator the result differs by an order or magnitude.

Any hint of the things we could check?

Many thanks

Please forget about this thread… We will open a new one more accurate and with an example.