One of the challenges with any kind of a quantitative system in football is that the interactions between players are extremely difficult to get a handle on. In the simplest terms, when Matthew Stafford throws the ball to Calvin Johnson, how much credit should Stafford get for the play and how much credit should Johnson get? That’s as easy as it gets. Now what do you do when Stafford throws to Tony Sheffler who is open because the defense is primarily concerned with stopping Calvin Johnson? Who gets the credit for the play then? According to Advanced NFL Stats, the greatest season that Randy Moss has had since their numbers began was in 2000 when he added 0.16 Win Probability Added/game. During that same season ANS credits Duante Culpepper with 0.28 WPA/game. But I consider it more likely that Randy Moss had a larger effect on Culpepper’s value, than Culpepper did on Moss’. Culpepper hasn’t done anything without Moss, whereas Moss allowed all of Culpepper, Randall Cunningham, and Tom Brady to have career years throwing to him. So what are we to think then of the fact that Culpepper’s WPA exceeds Moss’ when Moss is probably having all sorts of impacts on the game which can’t be measured. This isn’t an argument against advanced stats or even against Brian Burke’s WPA formula. It’s just a continuation of my argument that there is more to a passing game than a franchise QB. Consider that for the 2006 season, ANS credits Tom Brady with .11 WPA/game. In 2007, his WPA jumps to .38 WPA/game. Consider also that for the 2008 season, ANS credits Matt Cassel with .18 WPA/game and then that number dives to –0.15 WPA/game when Cassel leaves the Patriots for the Chiefs. So the effect of adding and removing Randy Moss/Wes Welker was +0.27 in WPA/game for Tom Brady when he gained those WRs and it was –0.33 for Cassel when he lost those WRs! Also worth pointing out is that Matt Cassel with Moss/Welker was worth more in terms of WPA than Tom Brady before he had them. Probably an important point here is that I’m not arguing against the value of Tom Brady. This is a two way street. Consider the following graph which shows my fantasy scoring efficiency measure for Andre Johnson. You can see exactly when Matt Schaub got to Houston. But the relationship goes the other way too. Here is Matt Ryan’s fantasy scoring efficiency graph. You can clearly see when Julio Jones got to Atlanta. To go back to the Patriots and what happened when they added Randy Moss and Wes Welker, consider that according to Advanced NFL Stats, this would be the ordering of the passing games in terms of WPA during the 2006-2008 time period.
- Brady/Reche Caldwell/Jabar Gaffney
Obviously the optimal situation is a franchise QB with two wide receivers who can get open at will. But dumpster-fire-QB with two wide receivers who can get open at will ends up beating the franchise QB with two scrub WRs.