A discussion of the problem of double-counting in climate modeling, specifically when the same evidence is used to both calibrate a model and then confirm the adequacy of the results. Steele and Werndl turn to a Baysian approach to argue for a method of incremental confirmation, making double-counting unproblematic. For a response to this argument see Mathias Frisch’s 2015 paper “Predictivism and old evidence: a critical look at climate model tuning”.