Sonil: I love all the advanced stats, as well as how far back they go historically! I'm curious as to whether I'm understanding the stats correctly. When I see that Christian McCaffrey, for example, had a receiving DVOA of 386, can I make an apples-to-apples comparison to WRs DVOA and conclude that in addition to generating the second-most rushing value, he also generated the third most-receiving value (only Michael Thomas and Chris Godwin had more than 386 receiving DVOA)? Or is it the case that since the replacement level RB has lower receiving production than the replacement level WR, RBs' receiving DVOA is inflated compared to that of WRs so they can not be compared?
Aaron Schatz: The answer is the latter. Wide receivers are compared historically only to other wide receivers, running backs to other running backs. Also the opponent adjustments are based on what position the receiver is. So it’s not directly comparable. McCaffrey didn’t generate the third-most receiving value overall, but you can say he generated the third-most receiving value based on his position.
It’s a pretty big difference. If we change McCaffrey to code him as a wide receiver instead of a running back, he drops from 386 DYAR to 102 DYAR! He would actually have a below-average DVOA as a wide receiver. This makes sense if you compare him to teammate D.J. Moore (3.2% DVOA). Moore had 1,209 yards with 64 first downs or touchdowns on 136 targets. McCaffrey had more targets (142) but had 1,005 yards with 58 first downs or touchdowns. So if we compared them with the same baselines, it makes sense that Moore comes out a little bit higher than McCaffrey does. The expectations for wide receivers are much higher than the expectations for running backs in the passing game.