Hi,
Extremely sorry for the very late reply. Is this related to your post in January?
I am curious whether the two samples that you want to discriminate have a large spread in event weights, what normalisation of the events you do and whether you still use BDTG.
Cheers,
Kim