Best Fit ? Estimated distance to minimum

Roofit and RooStats
1

HI,

I’m trying to (ML) fit my data with a crystal ball function. Every time I change the initial value of one of the parameters I get a completely different fit (with different number of signal events).
For all these fits I get the same minimized FCN value but with a different estimated distance to the minimum.
How can I know which one of my fits is better? Can I rely on the estimated distance to the minimum?
The fit with the smallest distance would be the best?

Cheers,
Simon

2

Hi,

It is really strange that you get different solutions for the same FCN value. Are you sure that the implementation of your FCN is correct ?
I would try to scan it as function of one or two parameters

Cheers

Lorenzo

3

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.