RooStats: HypoTestInverter


I’m trying to calculate limits using HypoTestInverter. I’m doing this using an invariant mass histogram for several different signal mass points. My issue is that when I plot the observed CLb vs. signal normalisation curves with HypoTestInverterPlot for the various different mass points they come out different for each mass point. This isn’t what I want to be happening since the observed CLb should have no dependancy on the signal model.

My best wild guess is that the code is doing something like restricting the mass range of the data that is used to calculate obs CLb to be around the signal mass window.

I’m using:

double mu=1.0;//For example
HypoTestInverter calc(*hc);
HypoTestInverterResult * r = calc.GetInterval();

to get the HypoTestInverterResult for each value of the signal normalisation, for each signal mass point and then use Add to combine them before plotting.

Do you have any idea what’s going on? Is there some default to restrict the mass range that I can override?


Ignore that post. Turns out it was just me not understanding that CLb is calculated depending on the mu value of the test statistic.