Playmaker Score: Kevin White and DeVante Parker Revisited

Since we unveiled the Playmaker Score projections for 2015 wide receivers, we've received a number of questions, whether they be through e-mail, Twitter, or the comment threads both here and on ESPN.com. Most of the questions revolve around two players, Kevin White and DeVante Parker, so we asked Nathan Forster to dig into those two situations a little more.

With Kevin White, the questions revolve around the fact that his status as a JUCO transfer might explain why it took him until his senior year to put up phenomenal numbers at West Virginia. Readers have asked how White would compare if we only looked at other JUCO transfers, and how JUCO transfers tend to do in the NFL compared to similar prospects who only attended a four-year school. Here's Nathan's response.

As described in the Playmaker article, Playmaker puts a large premium on a receiver leaving college as an underclassman. This could arguably result in an unfair penalty for players who spent their first two years of eligibility in junior college, who have only one year to build up tape strong enough to make it worthwhile to enter the draft early. Unsurprisingly, it is relatively rare for a JUCO transfer wide receiver to enter the NFL draft early. According to our Playmaker database, only two players have pulled it off since 1996: Robert Ferguson in 2001 and Cordarrelle Patterson in 2012.

That said, the evidence available does not suggest that Playmaker's decision to label former JUCO players as "seniors" rather than "underclassmen" results in an underprojection to any significant degree. There have been 44 former JUCO wide receivers drafted since 1996. The average error of these Playmaker Projections is an underprojection of 18.9 yards per NFL season. If these players are all re-coded and treated as underclassmen, their Playmaker Projections would overproject them by an average of 78.1 yards per NFL season. In other words, treating these players as seniors rather than underclassmen reduces the average error of the projections for these players by just over 75 percent.

However, if we are really skeptical we can limit our dataset to highly drafted JUCO transfers only. I took a sample of only those JUCO prospects who were projected to be drafted in the first three rounds by NFL Draft Scouts. Playmaker Projection underprojects that sample by 38.2 yards, a slightly larger amount. However, only 15 JUCO wide receivers in the database meet these criteria, and the effects may be simply noise. Moreover, even if we assumed that the signal from this analysis was real, it would be of only marginal help to White. If we were to revise White's projection upwards from 406 to 444 yards, he would still have a worse projection than Jaelen Strong, Sammie Coates, Breshad Perriman, and Nelson Agholor, all of whom can be had at a much lower price than White. Of course, White's Playmaker Projection could be wrong for a whole host of reasons, but it is extremely hard to make the case for drafting White over his peers based on analytics.

Questions about DeVante Parker are almost always about the fact that he was injured for half of his senior year. He finished the season with 855 yards and 5 touchdowns, and his Playmaker Projection is currently 404; what would his projection be if he had finished his senior season with 1,710 yards and 10 touchdowns instead?

Parker's projection comes out low for an entirely different reason than White's: Parker has a goofy history of production that includes three years of less than NFL-caliber play and one half-year of amazing play. Playmaker comes out low on Parker, however, because it is not designed to scale the production of a wide receiver to account for games missed. Thus, Parker's 855 yards in six games in 2014 are treated exactly the same as if he had 855 yards in a full 13-game season.

This design is purposeful, because projections based on production scaled to games played would have been less accurate historically. Although this may seem counterintuitive, in practice it negatively affects few players because Playmaker uses only a wide receiver's "peak" season. Thus, wide receiver prospects are not penalized for bad seasons, and are given a mulligan for any poor performance so long as they manage to string together one full college season that makes them look like a star. The best example of this is Dez Bryant, who had 323 total receiving yards in 2009 as a junior (in only three games), but his Playmaker is unaffected because he had an amazing sophomore season.

However, just because a lack of scaling would have worked in the past does not mean that it will always be the right approach in the future, so it would be worthwhile to see where Parker's projection would be if with a reasonable estimate of what his performance would have been over a full 13 game season.

To create a "hypothetical" Parker, I took his average receiving yards and touchdowns per game over his final two seasons, multiplied that number by seven (to account for the seven games missed) and added it to his actual numbers over six games. To me, this seemed to strike an appropriate balance between accounting for his improved recent performance and assuming some regression to previous levels. This accounting gives the hypothetical Parker 1,710 receiving yards and 11.2 touchdowns in 13 games. I also adjusted Louisville's team attempts so not to punish Parker because Louisville passed more, on average, in the games in which Parker did not play.

These adjustments, unsurprisingly, would improve Parker's projection significantly. Indeed, if Parker turns out to be a smashing success, we may need to incorporate a similar adjustment into the "official" version of Playmaker. However, even the hypothetical Parker would be projected to underperform his draft position. The hypothetical Parker would have a Playmaker Score of 68.6 percent, as opposed to the average Playmaker Score of 80.5 percent for a first-round pick. Indeed, even if his senior stats were simply pro-rated, assuming no regression to the mean, Parker's Playmaker Score would still be a tick lower than that of the average first-round selection's.

Again, as was the case with White, Playmaker may in fact prove to be wrong about Parker for a number of reasons. However, the question before evaluators is refreshingly simple: is Parker worth his projected draft position? From an analytics standpoint, that answer is "no."

Comments

There are no comments yet.