tW: theory covmats in ATLAS and CMS are different

[jacoterh]

The theory covariance matrix in tW is not the same for ATLAS and CMS, even though they share the same best_sm, meaning that the prediction has been obtained with the same calculation. For reference, see for instance

• https://github.com/LHCfitNikhef/smefit_database/blob/main/theory/CMS_tW_8TeV_inc.json
• https://github.com/LHCfitNikhef/smefit_database/blob/main/theory/ATLAS_tW_8TeV_inc.json

Action item is to ask this to the person who computed the theory covmat at the time of smefit2.0 (maybe @enocera remembers?)