Atlas 2015 i1394865 norm

I think the normalisation for this routine was messed up, and is fixed by this change. The plots are in TeV, and the normalisation attempts to deal with that, but it divides by TeV when I think it ought to multiply. It looks like this was introduced when the REF data was synced here: a77eaecf As far as I can tell by logic and looking at Contur plots this fix is correct but I am a bit nervous since I don't have a SM sample to validate with and I'm not sure how this would have passed validation if I'm right. So, I'd be grateful if someone else could sanity check...