Ravens Draft Parables: How to draft a wide receiver (Part 1)
In this series of posts I will combine my love of analytics, the NFL draft and the Ravens to tell a story, usually a cautionary tale, a warning, a plea or just some added insight into the Ravens and their draft.
There is a contradiction at the heart of being both an avid watcher of game film on prospects and being a self-confessed data geek. One believes fundamentally in people, not just the players you are watching, but more so, your own eyes and the eyes of other like-minded draft nerds. While the other says that those eyes are human and they are liable to a good old, self-inflicted hoodwinking – that the numbers tell a different story when it comes to your latest draft crush/next big thing prediction. This, as you will see Ravens Flock, is a sentence about Breshad Perriman.
At the risk of sounding like an older gentleman complaining about the state of the world, there are countless examples of folk who too quickly trust the provenance of a particular source – don’t question its inherent bias, or themselves, and don’t seek out the best version of the truth. They blindly follow either herd wisdom or information written by those with questionable motives. It was my usual desire to avoid this trope that led me to a more trivial question for those not part of our draft-loving community – how did I not have Justin Jefferson as my WR1 last year?
Let me draw the line for you from not questioning yourself to underrating the former LSU standout.
Much has been much written about the rookie phenom and his first season in the NFL, so I won’t rehash the point too much here but, in addition to breaking the rookie receiving yards record and keeping company with the likes of Diggs, Hopkins, Ridley, Adams and Hill, even the more sophisticated Next Gen Stats shows Jefferson in similar profile to his dominant contemporaries. The only slight differences being average yards of separation, which was a touch below some of those names and yards after the catch (YAC) which was a particular area of dominance for Jefferson – he was top 5 in yards after the catch and average above the expected yards after the catch – meaning he was creating a lot of yards on his own relative to the rest of the league. All of this confirmed when you watch his game.
This kind of statistical performance from a rookie in the NFL, must beg the following questions for the curious and ever-improving draft analyst - how, on God’s green earth, did the NFL take three receivers before this man and how did I have two receivers rated above him? And most importantly for this article – did the available data give us some kind of warning sign to expect this level of production? Granted, this wasn’t a Tom Brady-sized overlook but this guy should have been a top five pick and he was a consensus 20-30 type player and that was where he ended up being taken.
For NFL team success and success in draft analysis – predicting future NFL performance from past college performance is vital. So we know that watching the tape led most people to underrate Justin Jefferson, so was there anything predictive from other data about him?
Depressingly for those of us that got it wrong and who like to challenge human opinion with facts and figures, every possible publicly-available statistic from Jefferson’s final college season would lead you to believe that Jefferson was destined for greatness. His 2019 college season was Dominant (capitalisation deliberate). The data was screaming at you to not ignore this guy. Number one in receptions and in conversion of targets to receptions, number 3 in yards, number two in touchdowns and QB rating when targeted, even top five in yards after the catch. Yet I, and much of the draft community rated him as the third best wideout in the class. The NFL took him as the fourth best receiver after Jalen Reagor. It’s too early to judge many of the receiver’s in last year’s class but I think we can all agree that Jefferson is generationally good and he gave us the college production warning.
He did however end up below average in one particular area – yards per reception, in which he was the 54th best receiver in the nation in 2019. This stuck out so oddly amidst his absolute supremacy over other statistical categories that I dug deeper to see what statistical analysis of other classes might tell us about this anomaly and I found that underperformance in this data point was actually an important indicator of potential future success along with two other things.
It appears that college dominance of Justin Jefferson’s kind can be predictive of future performance if…
1. The receiver’s production is while attending a power 5 program
AND
2. They do not have a comparatively (to other receivers in college football) high yards per reception
OR
3. They do not have a comparatively (to other receivers in college football) high number of touchdowns
This seems odd that we require underperformance statistically to predict future success so let me explain why I think the last two rather simple statistical categories are important qualifiers for potential future success. We are all victims of our own unconscious bias in pursuit of greatness, and we all have a propensity for belief in our own opinions above all others. However, everyone who tries to evaluate players in projection to the NFL, from team General Managers all the way to me, sat on YouTube and searching for All-22 tape to conduct my own analysis, can all fall victim to the same simple problem. We all, at one point, fell in love with the game of football. So we all, at any point, can be blinded simply by players making great plays. If you find yourself embracing a receiver prospect either by watching them with your own eyes or being led to them by statistical dominance, use two simple statistics to challenge yourself and make sure you aren’t over-rating them because of the awe and wonder of either a high concentration of big plays, borne out by yards per reception, or a high concentration of touchdowns.
Oh, how I wish the Ravens challenged themselves with this when they drafted Breshad Perriman in the 1st round all those years ago. I know there were many reasons for his demise in Baltimore, not least John Harbaugh’s frustration with the all-time slowest healing PCL injury, but he was also number one in the nation in yards per reception on receivers with more than 87 targets that year. Ironically for me, if you accept receivers with less targets than this into the list, Devin Smith comes out on top of the yards per reception stakes – he was the guy I wished the Ravens had drafted and even more of a bust. So I can’t criticise too hard, I haven’t checked myself with this before.
There were many lessons and stories to be told from my wide receiver data journey but before I become the white rabbit and lead you all, as a collective Alice down the rabbit hole in part 2, let me leave you a summary of the main lesson for today.
Dominance at the receiver position in a power five conference, the likes of which we saw from Justin Jefferson at the college level cannot be ignored.
This really, is a lesson about Devonta Smith and to a lesser extent Ja’marr Chase.
Justin Jefferson was a top 10, in his case top five, contributor in eight simple statistical categories including receptions, reception % (from targets), receiving yards, receiving touchdowns, YAC, 1st downs, avoided tackles, and QB rating when targeted. He also had the all important yards per reception anomaly. Devonta Smith matches him with eight top five finishes and the yards per reception anomaly. Devonta Smith is an elite talent, do not fade him in any way. Ja’marr Chase’s 2019 season matches Jefferson’s in the most part but wasn’t quite as dominant and did have a high yards per reception and high number of touchdowns. While he didn’t have the all-important anomaly, he did look like no other receiver statistically due to his also high YAC, meaning he did do a lot of the work himself to create that high number of yards per reception. But Chase vs Smith will be an interesting test balloon for this hypothesis with Smith likely to outperform Chase based on the findings.
Finally I have two key disclaimers to all this – firstly, the majority of the really good receivers in the league did not signal their future ability with standout college statistical production. This is all about not ignoring dominance rather than finding gems. And secondly, we of course, cannot simply trust the data – the bare facts will never tell you the whole story. We must go back to the game film and spot the skills and traits that we know translate well to the NFL. The data can help us along the way a bit, challenge us, or confirm what we think we know - when we know what to look for, hopefully you do now.