giacomomagni
commented
1 year ago Hi @arossia94, yes I agree we should store the kinematics variable somewhere, I think the best way would be to include them in the experimental data files. Note this is also needed in order to produce fits with cuts. However with @tgiani we realised it is not straightforward, since there is a large variety of processes each with different kinematics. A also discussed here we are now implementing a sanity check on the SM value, I think we can try to tackle this problem there (or at least give a first try).
In the meantime I'm quite confident that all the dataset (exp and theory) are sorted with the same order if this was not the case than we were just fool ourself... But as said it would be great to see the kinematics of each point!
There seems to be no way to identify to what corresponds each array element inside the theory table of a given dataset beyond comparing numbers. For example, in the ATLAS_SSinc_RunII theory table, how do I know which element corresponds to each signal strength?
That's kind of the worst case because the SM prediction is 1.0 by definition and hence, I should check the dependence on the WCs to identify each signal strength. I could identify them by checking with the corresponding data, but that requires trusting that the ordering was the same in both.