Data mining is indeed the problem. Finding the length of every TD pass for a given QB is easy, but finding the length of the rushing TD’s his teams scored while he was on the field is all but impossible without going through each boxscore by hand.

Frankly, the only reason I kept it this simple was in the interest of time. I used the .596 multiplyer to put offensive touchdowns on the same scale as TD passes, in the hopes of making the comparisons more intuitive, but you’re right that it isn’t really necessary. Also, play-by-play only goes back to 1999 so the historical perspective would not have been possible.

I like your suggestion, and actually was contemplating something very similar (adjusting for goal-to-go situations). Long TD passes should almost always be credited to the QB, whereas a one yard TD has far more to do with playcalling. Those factors definitely need to be sorted out in order to have the most accurate results. Perhaps a future study is in order…

Thanks for reading and commenting!