This question has been asked in the past (See Refs [1] , [2], [3]) but since some years passed I want to make sure we understand the correct way of dividing two TEfficiency objects.
Assuming we have two uncorrelated TEfficiency objects, A_eff and B_eff , should we
I believe the answers to the previous ROOT questions are still valid even though a few years have passed. We also keep our documentation up to date, so I believe you can also follow what is written here. Nonetheless, let me tag @moneta for some more detailed insights.
I read the documentation of TEfficiency and there is no report of taking the ratio of two such objects. So there needs to be some conversion of them to TH1 or graphs before proceeding with the division. At least that’s what I understand from Reference 3.
As there is not a written example in the forum, I took the liberty of re-opening the issue.
My answer post few years ago, Taking Ratio of TEfficiency Distributions - #4 by awisecar , it is still valid.
We don’t have way to compute correct uncertainty in ROOT for this case. You can maybe use an approximation, which will be more or less valid depending on your data values.