Best fit result and the minimum of NLL show difference

Roofit and RooStats
1

Dear expert,

Hello, I’m trying to 2D fitting and using the fitting result.

I would like to get 90% CL Upper limit of signal yield(nsig).

I thought that the fit results about nll value and signal yield

and the minimum value of nll and the nsig value at that minimum point should be same.

but it was not… I don’t know why…

the fit result shows that nsig is about 2 +/- 18.

but the nsig value at minimum point of nll is about 65.

the attached file is my code for fitting and estimating Upper Limit.

Any help would be helpful to me and would be appreciated.

Cheers
Chanho

excl_genD2nunu_2Dfit.C (10.2 KB)

2

Hi,

Is it possible to reduce your example to a minimal reproducer, provide the input files to us while the experts have a look @moneta @jonas ?

Cheers,
Danilo

3

Thank you for replying.

While investigating more,

I found that I used different data for calculating likelihood function…

So, I fixed that kind of mistakes,

and now it shows reasonable nsig value at the nll minimum point.

Cheers,
Chanho