Full disclosure, I put this up on r/NFL. I apologize if the cross post is against the rules. I figure if it's worth anything it's worth something in both places. But again I'm sorry if you've seen it twice.
I don’t honestly know what the sophomore slump is supposed to be. I know it’s alliterative and that scares me because we naturally trust attractive aphorisms - here there be PDFs. So when I hear Sophomore Slump used it grates on me - and not just because Russell Wilson will only ever get better but because I think it represents the laziest part of our minds. The part that sees assonance or rhyme and thinks "oh, yeah, that’s probably right then."
Actually it’s even worse than that. Take "fortune favors the fool" as an example. It’s predictive. It makes a tangible claim that can be proven or disproven. You might say "fortune favors the fool, just look at
Pac Man Adam Jones." Then I’d say "Oh, okay. Look at this trend line I just drew on your face showing that education correlates strongly to wages." Point made.
Sophomore slump isn’t predictive on its face. On its face it’s adjectival. It just describes a second effort worse than one’s first. I can’t deny that The Arcade Fire or Matt Ryan had sophomore slumps, those are both objective facts based in data. So when someone uses it to predict a general trend it is worse than "fortune favors the fool" because it’s less clear how to help them
feel bad for being stupid adjust their world view. In particular because the phrase practically begs for football fans to get in anecdote fights. A pastime only slightly more popular than watching actual games and only slightly less likely to result in change.
And people do use sophomore slump to predict general trends. People use it because it’s narrative. It makes a story in a way regression to the mean, interception luck, or developing old player skills simply can’t. It let’s us talk about whether Cam Newton is pouty (no) or if Matt Ryan has eaten at the Flying Biscuit Cafe one too many times in the off season (you literally can’t eat there enough).
BUT IT’S NOT REAL.
Let me be more specific: There is no general trend for sophomore QBs to perform worse than in their rookie season. There are no specific trends for sophomore QBs to perform worse than in their rookie season. There is no reason for people to keep using this stupid loaded phrase prospectively.
There are 47 QBs between 1970 and today who have had >223 pass attempts their rookie and sophomore seasons. I have included every one of those QBs in my data set. Note that this includes guys like Warren Moon, Jeff Garcia, and Jim Kelly who played pro football in places other than the NFL before their rookie season because I’m not going to wiki each of the guys who played before I was born.
I used ANY/A (Adjusted Net Yards per Pass) as my primary measure of performance. Generally I don’t like QB stats that include interceptions since they don’t appear to be a repeatable skill but I’m using ANY/A instead of NY/A here because what if all sophomore QBs throw more interceptions than they did as rookies?
The Results and Analysis:
The change in ANY/A (I’ll call it ΔANY/A) between rookie and sophomore campaigns ranges from -2.11 for Dennis Shaw (who won the 1970 OROY) to 3.25 for Josh Freeman (did not win OROY in 2009). The average ΔANY/A was .67, an improvement equivalent to Drew Bledsoe or Christian Ponder between their sophomore and rookie seasons. 12 out of the 47 QBs got worse and one stayed exactly the same (cool). There is no general sophomore slump.
The correlation coefficient between rookie season ANY/A and ΔANY/A is negative but it’s also quite small (-.29). This could mean that only good rookie QBs have the sophomore slump but after eyeing the plot of the data it is clear that the correlation exists because bad rookie QBs can’t get worse or aren’t given a chance to. Steve DeBerg and Joe Pisarcik both had an ANY/A below 2 their rookie years. Those two seasons are fully 1/6 of the seasons in which a QB has thrown for less than 2 ANY/A. Because the bottom left quadrant of the graph, which would have bad rookies who played worse as sophomores, is empty the correlation coefficient is necessarily negative.
There is essentially no correlation between age and ΔANY/A (-.14). I checked on this to see if guys like Moon and Garcia were throwing off results. In the end I think it was more or less a waste of my time but since I did it I’ll throw it up here.
People need to stop asking who is going to beat the sophomore slump. Obviously it’s going to be DangeRuss.