I would like to use the Chi2Test method of TH1, and possibly the KolgomorovTest method, to compare two TProfile plots. The problem I have is that in some regions of the plots, the associated errors are not reliable (they are too small - binomial errors of proportion issue), and so I wish to discount these from the test. Is it possible to define a region over which the test will be applied? Failing that, how can I override the Bin error to just set it to a massive value so that it doesn’t contribute to the Chi2? I tried used “SetBinError” but this doesn’t seem to have an affect on TProfiles?
Cheers,
Will