-
Notifications
You must be signed in to change notification settings - Fork 3
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Sum of weights for non-EFT process #27
Comments
@btovar do you have any thoughts on this? Just changing the default type to |
@bryates I think that was the idea, to record it separately from that double axis. As I understood it, |
Right, we don't really need it for the actual EFT histograms (sine it would be a quartic function), but we'll need it for the SM terms that are subtracted off later one. Do you know of a coinvent way to add another axis, or would it be easier (but maybe less efficient) to just save another histogram with the weights^2? |
I think it would be much easier to have another histogram with the weights. The way the new histograms work make it hard to have special treatment of some axis. |
Ok thanks. I was starting to suspect that after trying to hack something in. |
Forcing the storage type to
Double
topcoffea/topcoffea/modules/histEFT.py
Line 101 in 4c53b0b
means we lose the sum of weights, even in the non-EFT samples. The data is simply
sqrt(N)
, but the nonprompt leptons have an MC component subtracted offhttps://github.com/TopEFT/topeft/blob/3ae0f05890a129aaa226efeaec949d13e1778468/topeft/modules/dataDrivenEstimation.py#L115-L129.
This means the nonprompt templates have no uncertainties, but we still ask combine to compute the MC stat errors. This is likely causing the remaining discrepancy in the limits when compared to the master branch on the TopEFT repo.
The text was updated successfully, but these errors were encountered: