“Imagine that we want to test the claim that a special diet – let’s call it D – that consists of consuming 10,000 calories a day without any exercise will lead to weight loss of 50 pounds in 5 weeks. To make such a bold claim you would expect us to examine the evidence by comparing the recorded weight of people before and after being on the diet. But suppose instead our ‘evidence’ is nothing other than following mathematical model: Weight loss after n weeks in pounds = n x efficiency_D where efficiency_D is the average weekly weight loss of diet D. Suppose further that we assume: efficiency_D = 10 Then, with these assumptions we compute Weight loss after 5 weeks in pounds = 5 x 10 = 50 QED! We can now trust the claim that the special diet D does indeed lead to weight loss of 50 pounds in 5 weeks and we can believe this with full certainty (hand on heart). Would you be convinced by this argument?