Recently I wrote this article on Cordarrelle Patterson and Justin Hunter, arguing that (despite his subpar market share of receiving yards) Patterson is the superior prospect and more deserving of a (late) first-round pick in rookie drafts. RotoViz’s Shawn Siegele responded with this fantastic rebuttal, restating and adding to his case against Patterson. While I agree with almost everything Shawn said in his article, I wish, like the empire, to strike back!—well, at least to address some of the legitimate concerns Shawn raised when pointing out the dark side of Patterson as an NFL prospect. In doing so, I understand that I risk becoming to Patterson what Frank was last year to Marvin McNutt (an obsessive advocate), but nevertheless I shall continue, because what is at stake in an evaluation of Patterson, as implied in my original piece, is not just Patterson’s future prospects but also the question of how best to go about evaluating WRs entering the NFL.
In responding to Shawn’s article, I hope to show that, in all his “Sithiness,” Patterson is more Vader than he is Emperor. (And I am clearly more nerd than I am not.) What I mean to say is this: Despite his flaws, Patterson as a prospect can be redeemed. If evaluators place draft prospects into one of two columns, “draftable” and “undraftable,” I believe that we can bring Patterson back to the good side. Even though he sports the receiving market share of an undraftable player, he is unreservedly draftable.
In fact, I’ll start by discussing and clarifying my ideas on Patterson’s potential draft position. In his conclusion Shawn wrote this: “Matthew suggested that Patterson will go late in the first round of rookie drafts.” I may have inadvertently implied this in my piece, but here is what I actually wrote: “Patterson’s still worth a draft pick if he reaches a certain spot in rookie drafts. [. . .] And, if you find yourself in a rookie draft with no first-round-worthy RBs on the board, such as Knile Davis, and if Geno Smith and a couple of WRs have already been selected, then you could probably do a lot worse than taking Cordarelle Patterson with your (presumably low) first-round pick.”
To clarify, I do not think Patterson will be available late in the first round in many (or most) rookie drafts, but I can see scenarios in which he slips—especially in strong dynasty leagues, and I assume (perhaps wrongly) that if someone is reading our articles (s)he is probably participating in some leagues with strong players, and Patterson seems to be the kind of prospect that strong dynasty players with high picks may choose to avoid. For instance, I could see three or four RBs being chosen with the first six picks, because teams drafting high often need or covet RBs. And, after what the 2012 rookie QBs did, I could see Geno Smith going in the first six picks of some rookie drafts (click here to read Shawn’s killer piece on the QB), especially when one considers his high Lewin Career Forecast of 2064 DYAR. And then potentially Tavon Austin, Keenan Allen, and DeAndre Hopkins could all be chosen before Patterson—again, depending on the league.
In other words, I can see how Patterson may slip slightly in some rookie drafts, in the range of Picks 5-9. In most leagues, he won’t, but in “good leagues” he could. Of note is that Shawn and I both agree on what to do if he is available late in the first round. To quote Shawn, “If that’s the case, you should snap him up immediately.” I concur. Late in the first round, he presents nice value, as WRs taken in the top-half of the first round have strong chances of producing usable fantasy seasons.
“A 6’2” Receiver Playing the Percy Harvin/Randall Cobb Role”
Also, I would be slightly remiss if I didn’t draw your attention to this article by Nick Whalen, posted the same day as Shawn’s, which highlights Patterson’s elusiveness. (And I should add that the brand of “mathematical scouting” Whalen exhibits in his article is exactly the type tape-grinding analysis I was advocating in this utopic piece. Job well done.) What Whalen’s piece suggests to me is that, if used in a certain fashion, Patterson has the potential to be a special NFL player. In his piece Shawn said that it would “be strange to see a 6’2” receiver playing the Percy Harvin/Randall Cobb role,” and I agree—but aren’t Harvin and Cobb two of the coveted dynasty receivers in the game? If Patterson gets in an offense dedicated to giving him the rock quickly and in space, he could be better than Harvin and Cobb since he is not only faster than Cobb and just as fast as Harvin (judging by their unofficial combine 40 times) but also more powerful than both.
And, if a team takes up the Sisyphusean endeavor by drafting Patterson, it will probably do so with the idea of using him as not a traditional receiver but rather a Harvin-esque offensive force. Even without a “draftable” Dominator Rating of 0.25, Patterson is imminently draftable because he will likely not be used as Darrius Heyward-Bey. Although he is closer to DHB in size, Patterson is closer to Harvin and Cobb in style. After four years in the league, Harvin has proven himself to be the most productive WR of the 2009 Draft, which featured a first round in which—get this—six WRs were selected, all of whom have produced at least one top-30 season. If you don’t know how rare it is for the top six receivers from one year all to have top-30 seasons, much less for all those WRs to be first-round players, then I’ll be happy to tell you—it’s unprecedented. From this perspective, the first-round WRs of 2009 are historically unrivaled as a cohort, and the non-traditional Harvin has been the best.
That’s the reason I like Patterson more than other RotoViz writers do: I think he will be used like a big Percy. He may not be as effective as Harvin, but at least we can expect him to be used in a similar fashion, and when used that way in college he excelled. And, for what it’s worth, let me be clear about this—I don’t like Patterson that much; I just like him more than the other RotoViz studisticians do. I’m not saying he’s A.J. or Julio. What I am saying, though, is that I wouldn’t be surprised if Patterson put up a Harvin-esque or Santana-esque top-3 finish in his fourth year.
In my article I pointed out that many good NFL receivers had at least one season of 75 rushing yards in college—and I noticed, but forgot to point out, that most of these receivers also had solid DR scores as well. Rightly, Shawn pointed this out, and I am glad he did. He is absolutely right. Most of these good rushing WRs also dominated as receivers. Of note, though, is this intriguing (and I dare say predictive) fact: Many of the rushing WRs with subpar DR scores (below 25%) still had strong NFL production. Who are those guys? Santana Moss (22.4%), Harvin (24.5%), and Wes Welker (17.4%). And Mike Wallace (26.6%) and DeSean Jackson (26.5%) were barely past the 25% threshold that Shawn says a WR requires to be draftable.
Clearly, not all players with DRs this low are draftable, and no system will perfectly include all productive players while excluding all non-productive players, but I imagine that a system could be created that both highlighted Santana, Percy, and Wes and excluded the low-DR receivers without realistic prospects for NFL production. Specifically, I think that such a system could make use of raw stats and rushing yards. Please note that Welker’s DR is almost exactly the same as Patterson’s (17.5%). Is a 17.5% DR horrible. You bet it is, and you shouldn’t even think about touching almost all guys with that kind of rancid DR. But Welker and Patterson? Yes, you should think about it. Why? Because both guys had at least one season of 75 rushing yards, and both guys had final collegiate seasons of 1000 scrimmage yards and 10 all-purpose scores. The rushing yards matter, as do the raw stats. Those two factors would eliminate almost all subpar DR receivers while preserving a small core group of players with the legitimate pro potential normally associated with players possessing better DRs.
Here’s another way of looking at rushing WRs: The skill that enables them to be good runners also enables them to be good receivers (so the hypothesis goes), and so most of them exhibit good DRs, but some of them do not. The players who don’t, however, still possess the potential to have high DRs—they are good enough to have high DRs—but, for whatever reason, they simply don’t. In a manner of speaking, the high DR is latent within the low-DR rushing WR and has simply not yet expressed itself. Why do I say this? Again, because the skill that enables WRs to be good rushers enables them to be good receivers, and not vice versa. For this cohort, the above-average correlation with NFL success is not linked to high DR. Rather, it is linked specifically to the skill that enables good rushing, a high DR, and sometimes both. That skill is what we really seek to find when analyzing WRs, and DR is a great tool for identifying it in many players. What I am suggesting is that rushing yards are also indicative of the presence of that skill in WRs.
One more analogy: If Cordarrelle Patterson were an investment with a price chart that one could trace over time, “rushing yards” would be one of the technical tools one could use in the analytical process, and the presence of rushing yards could be taken as a leading indicator, suggesting a potential breakout in the future. Dominator Rating would also be one of those tools, one of the more trusted and obvious, and so if Patterson had a high DR more people would recognize him as a sound investment and the price to acquire him would rise. This part is important—Patterson’s failure to register a high DR represents a buying opportunity. His lack of a high DR does not necessarily (although it could) mean that he won’t have a breakout in the future, since DR is not the only analytical tool, but it does mean that his potential for a breakout—signaled by his rushing yards—will not be as obvious to some dynasty players.
The most lucrative investments are sometimes those in which the potential breakouts are signaled not by widely known analytical tools but by the more underused and neglected. Patterson may not end up being a highly valuable investment in dynasty drafts (dependent on the pick used to acquire him), but I undoubtedly think that he should be considered in a class with Harvin, Welker, and Santana Moss—guys who, as NFL players, outperformed their collegiate DRs. In effect, Patterson’s raw stats and the existence of his rushing yards should be given more relative weight. Or, put differently, he should be treated as if his DR were higher. Or, once more, with Patterson perhaps we should ignore the concept of DR altogether, since the other leading indicators are those that more precisely seem to prove predictive for his type of asset class—the Harvin-Welker-Santana peripheral/slot growth stock.
“Advanced Metrics Do Matter”
I almost hate to suggest the very idea of ignoring DR when considering Patterson, because, as Shawn says in his article, “advanced metrics do matter.” I am starting to think, however, that perhaps different types of players should be considered using different criteria.
Here is a case in point. In arguing that advanced metrics matter Shawn compares the (admittedly cherry-picked) raw stats with the advanced metrics of Demaryius Thomas and Titus Young. The raw stats are almost identical, but the Dominator Ratings are vastly different, as are their Height-adjusted Speed Scores.
I’m glad that Shawn made this comparison, because it may help me clarify some of my points. What I wanted to suggest most in my original piece on Patterson was the utility of employing both raw stats and advanced metrics. In my system of analyzing WRs, Titus Young was undraftable coming out of college because of his physical profile—he was too slow for a guy of his weight and height. In other words, for me, Young’s stats—raw or advanced—don’t even matter because the advanced metrics regarding his athleticism eliminate him from consideration. Demaryius Thomas, though, was draftable because of his positive physical metrics, sufficient raw stats, and—more to the point—the round in which he was drafted.
Young was a second-round WR, and I judge that type of player differently than a first-round WR. That may be unfair, but, as I am beginning to show in my series on first-round WRs and their historical performances and value, first-round WRs (at least some of them) are highly valuable commodities, and so the criteria that I use in evaluating them are less stringent than those I use for other WRs.
And what are these criteria for first-round WRs? As long as a WR has certain physical attributes (partially informed by advanced metrics), I have found that either 900 scrimmage yards or 9 all-purpose TDs in his last college season strongly correlates to a degree of NFL success. If a “physically qualified” first-round player hits either of these benchmarks, he is highly likely to have at least one top-30 season (and probably more such seasons) within his first four years. How correlated? Almost 90% of the cohort WRs. That’s pretty good, especially when one considers that about only 67% of all first-round WRs ever have a top-30 finish.
Furthermore, these benchmarks tend to divide quite neatly the players who become productive from the players who don’t. For the most part, the WRs who hit at least one of these statistical benchmarks have top-30 seasons, and the WRs who don’t hit either benchmark don’t have top-30 seasons. The system is not perfect, but it has proven incredibly useful in providing a quick answer to this question: Is this particular player draftable? If a first-round WR has the physical attributes I desire and either 900 scrimmage yards or 9 all-purpose scores (in at least or prorated to 12 games), then he is draftable; if not, probably not.
Where advanced statistical metrics would fit into this system (I believe) is in the ranking of the players who are already determined to be draftable via prior considerations of physical attributes and raw statistics. In this stepwise system, raw stats and athleticism would serve as the gatekeepers, used to differentiate (or classify) players, including and excluding WRs appropriately. Then, advanced statistical metrics would serve as the sorters, used to rank (or stratify) the cohort players.
Let me provide an example, based on this excellent consideration of Patterson in the context of historical first-round Productivity Scores by RotoViz’s Jon Moore. A Productivity Score is Moore’s market share metric, and the cutoff he has established for first-round predictiveness is 60—if one of the first-round WRs considered since 2005 has a 60 or greater, then he will probably have some NFL success. If he has a PS lower than 60, then he probably will not. Go ahead and take the time to click on the link and look at the article. It’s great stuff.
Without taking into account whether any of the players Moore considers actually pass the physical requirements I desire in WRs (in other words, the comparison is strictly between market share and raw stats), I think a comparison of his system with mine is instructive. Moore notes that 22 of the 28 first-round WRs since 2005have PSs of 60 of greater. Impressively, 16 of them (72.72%) have produced at least one top-30 season, with 4 players (18.18%) yet to do so (A.J. Jenkins, Michael Floyd, Kendall Wright, and Jonathan Baldwin) and 2 players (9.09%) who left the NFL without ever doing so (Mike Williams and Troy Williamson). Additionally, of the 6 players with PSs below 60, only 2 of them (33.33%) have produced at least one top-30 season (Percy Harvin and Mark Clayton), with 1 player (16.67%) yet to do so (Ted Ginn Jr.) and 3 players (50%) who left the NFL as unproductive receivers (Buster Davis, Anthony Gonzalez, and Matt Jones). Overall, market share alone proves quite predictive, and the cutoff point of 60 provides a nice line of division.
A system based on raw statistics, however, does just a little bit better—not enough to make one think that raw stats should entirely replace market share in one’s analytical process, but maybe enough to suggest that both could be used together. If one uses either 900 scrimmage yards or 9 total TDs as the cutoff, then almost everything stays the same, except that Percy Harvin, Mark Clayton, and Ted Ginn Jr. are now considered draftable and Troy Williamson, Darrius Heyward-Bey, Jonathan Baldwin are not. The effect is that 17 out of 22 drafted players (77.27%) produce a top-30 season, with 4 players (18.18%) still yet to do so (Jenkins, Floyd, Wright, and Ginn) and now only 1 player (4.55%) who failed to produce a top-30 season (Mike Williams). Further, of the 6 players who failed to hit either statistical benchmark, only 1 of them (16.67%) produced a top-30 season (Heyward-Bey), with 1player (16.67%) yet to do so (Jonathan Baldwin) and 4 players (66.67%) who finished their careers without a top-30 season (Davis, Gonzalez, Jones, and Williamson).
What does all of this mean? In sum, the system based on raw stats is, with this cohort at least, more predictive and efficient, moving one more productive player into the draftable category and one more unproductive player into the undraftable category. This may not seem important, but every analytical edge helps and when the players swapped are Harvin and Williamson the effect of that trade is significant. In fact, Percy v. Williamson is a microcosm of my entire perspective, of everything I have been attempting to communicate when talking about Patterson. Even without a “draftable market share,” Percy was draftable because of his raw stats, his rushing yards, and the manner in which he would be employed in the NFL. And, just as importantly, even with his impressive Productivity Score, Williamson wasn’t draftable because his raw stats were horrible—and a high Productivity Score paired with weak raw stats doesn’t mean that a player is undervalued; rather, it means that he has almost no value at all. Advanced metrics certainly matter, but so do raw stats. Indeed, in fantasy football, they’re all that matter.
I mentioned earlier that my criteria for non-first-round WRs are more stringent than the 900/9 raw-stat thresholds. How stringent? A WR in a round as early as the second must have at least 1000 scrimmage yards plus 10 all-purpose scores (in addition to meeting my standard physical requirements). I’ve found that lessening the rigidity of these statistical criteria results in inefficiency. And in the second round this system is still pretty efficient. In general, just under 40% of all second-round WRs ever produce a top-30 finish. My system, which considers raw stats and physical ability, bumps along at over a 70% rate.
The result is that I have decided not all WRs should be judged by the same criteria because they are not on the same probabilistic playing field. As a first-round WR, Patterson strikes me as a guy who should be judged by lesser standards, because historically lesser standards have been sufficient to separate the good first-round WRs from the bad—and, what’s more, Patterson not only surpasses the reduced standards for first-round WRs, but he meets my stricter standards set for all other rounds as well.
One downside to a system based primarily on raw stats is that it classifies but not stratifies. It says nothing about how high a player’s best positional ranking is likely to be or how many top-30 seasons he is likely to accrue, or the order in which WRs should be selected. While answering perhaps the most fundamental of dynasty questions (“Is this guy even worth drafting? Will he ever produce a startable season?”), it does little else. Here is where advanced statistical metrics could be used—to provide a basis for draft rankings. In effect, we could do almost the exact same draft analysis we already do, but we could also account (through attention to rushing yards and raw stats) for those few WRs with subpar market share who are still likely to have NFL success.
I believe Patterson is one of those players, and Shawn most likely does not. I am new enough to “fantasy writing” to know that I know comparatively little, and so if I had to bet on who would be correct about Patterson in five years, myself or Shawn, I would probably bet on Shawn. Still, I feel that the bullish argument for Patterson is worth making, because it points to a class of productive players overlooked by an approach grounded in market share. Even if I’m wrong about Patterson, I think I’m right about rushing WRs and the idea that players who differ in style and draft status sometimes should be judged by different metrics, and if this dialogue leads to the eventual creation of an efficient system that includes players like Harvin, Welker, and Santana then all this electronic ink on Patterson will have been worthwhile—maybe not for you the reader, but definitely for me and RotoViz.