In a recent article published in Synthese, philosopher of science Pekka Syrjänen asks “does a theory become better confirmed if it fits data that was not used in its construction versus if it was specifically designed to fit the data?”