update closure runcard to use new candidate methodology

Created by: wilsonmr

Also since the preprocessing slightly changed I generated a new single replica underlying law. Probably the old one was fine but I don't want there to be inconsistency.

I took the runcard from https://www.wiki.ed.ac.uk/display/nnpdfwiki/List+of+final+fits+for+NNPDF4.0+paper specifically NNPDF40_nnlo_as_0118_base and then I changed the datasets to the agreed 3.1-like dataset and the t0/fakepdf and the relevant closure flags. All in all should look pretty similar to old runcard but with smaller architecture etc.