RooMCStudy::PullPlot Problem with RooCBShape

[xiaotao]

Dear RooFit experts,

The mean value from pull distribution shifted from 0 when I used the RooCBShape as a pdf model, however it seemed normal when RooGuassian was used as a pdf model.

I tried a lot of ways to solve including fixed the tail parameters like alpha, n, unfotunately, it remains. Could you tell me how to solve the above problem. Thank you much for your help.

The code can be seen in the attachment. Thank you so much for your reply.toymc_test.py (1.7 KB)

RooCBShape was used as a pdf model.

RooGuassian was used as a pdf model.

[moneta]

Hi,

I think this is caused by a known problem (bias) of the binned fits, that they do not integrate the function in the bin. See https://sft.its.cern.ch/jira/browse/ROOT-3874

If you increase the number of bins (e.g. 1000) by adding m.setBins(10000) , I see this bias going to zero

Cheers

Lorenzo

[xiaotao]

Dear Lorenzo,

It’s so kind of you and Thank you so much for your solution, it works now by adding m.setBins(10000).

Best wishes,

xiaotao