In all practicality, you are fine with an estimate. You will never be spot on via formula anyhow, so will end up wasting a full test print, when you could have just done a test strip. Going from one standard print size to the next (say from 8x10 to 11x14), I would add one stop of exposure as a general adjustment. Going to 16x20? Add two stops versus the 8x10. This is making the quite general estimate that each standard size contains twice as much surface area as the next smallest. This is not exactly true, but it is close enough to make a good test strip. One thing that is true is that when going from 4x5 to 8x10, or from 8x10 to 16x20, you exactly quadruple the surface area, therefore you end up with only one quarter the light. This means that you can theoretically compensate by increasing the amount of light by a factor of four, also known as adding exactly two stops of exposure.