Writers of Pro Football Prospectus 2008

Most Recent FO Features

Injuries cover.jpg

» Injuries in the Post-Probable Era

Guest columnist Zachary O. Binney looks the effects of the removal of the "Probable" designation from the NFL's official injury reports.

02 Apr 2015

Playmaker Score 2015

Guest column by Nathan Forster

You can probably deduce whether your team had a good draft in 2014 if you can answer "yes" to one of the following two questions: 1) "Did your team draft Aaron Donald?" 2) "Did your team draft a wide receiver in the first round?" Last year's wide receiver class is only a year old, but it may be on track to be the best ever. Ten rookie wide receivers gained more than 500 yards receiving in their initial season. Three of those went over 1,000 yards. Although the success of these players is incredible, it can hardly be called surprising. Conventional wisdom anticipated that the 2014 draft class was chock full of wide receivers who were likely to be successful -- and so did Playmaker Score, Football Outsiders' system for projecting wide receivers

So can the 2015 draft class match its immediate predecessor's accomplishments? According to Playmaker Score, probably not. Although the 2015 class includes some strong prospects and a sprinkling of interesting sleepers, there are also a number of players with bust potential, rated higher by scouts than by Playmaker Score. It adds up to a potentially good, but not great, class of pass catchers.

Playmaker Score is based on a statistical analysis of all Division I wide receivers drafted in the years 1996-2012. Playmaker Score consists of the following elements, which are the factors that historically correlate to NFL success:

  • The wide receiver prospect's best or "peak" season for receiving yards per team attempt (i.e., a wide receiver with 1,000 receiving yards whose team passed 400 times would score a "2.50")
  • The wide receiver prospect's peak season for receiving touchdowns per team attempt
  • The difference between the prospect's peak season for receiving touchdowns per team and the prospect's most recent season for receiving touchdowns per team attempt (this factor is simply "0" for a player whose peak season was his most recent season)
  • The wide receiver's vertical jump from pre-draft workouts
  • A binary variable that rewards players who enter the draft as underclassmen and punishes those who exhaust their college eligibility
  • The wide receiver's college career yards per reception
  • The wide receiver's rushing attempts per game during their peak season for receiving yards per team attempt.

Playmaker Score has two outputs: "Playmaker Rating" and "Playmaker Projection." Playmaker Rating is the "purest" output for Playmaker Score: it is expressed as a percentage that measures how highly the player ranks historically based on the factors evaluated by Playmaker Score. For example, a player with a 75 percent Playmaker Rating scores more highly than 75 percent of wide receiver prospects drafted since 1996. Playmaker Projection is a more realistic measurement. Playmaker Projection acknowledges that a player with a first-round grade and a mediocre Playmaker Score is more likely to succeed than a seventh-rounder that Playmaker Score loves. Thus, in addition to the Playmaker Score factors, Playmaker Projection also incorporates a transformed variable based on the player's projected draft position from NFLDraftScout.com.

Here are the Playmaker Scores for the top wide receiver prospects available in the 2015 NFL Draft.

Amari Cooper, Alabama

Playmaker Projection: 643 Yards
Playmaker Score: 95.4%

Amari Cooper is head-and-shoulders the best wide receiver prospect in this class, and he hits most of Playmaker's buttons. Cooper had an impressive junior campaign, gaining 1,727 receiving yards and 16 touchdowns. Moreover, there is also nothing about Cooper's situation that gives reason to doubt Playmaker's numbers. Unlike many wide receiver prospects in this year's draft, Cooper did not play in the spread, and he faced some of the toughest college defenses in the country. The one knock on Cooper is that his combine numbers were just OK, and his vertical leap of 32.5 inches is below average for a drafted wide receiver. However, the correlation between the combine workouts and wide receiver success is surprisingly weak.

Historically, Cooper probably best compares to Indianapolis Colts great Marvin Harrison as a prospect; Cooper was not quite as dominant as Harrison was at Syracuse, but the mid-'90s Big East wasn't a match for today's SEC, either.

Jaelen Strong, Arizona State

Playmaker Projection: 514 Yards
Playmaker Score: 80.7%

Jaelen Strong's No. 2 spot on the Playmaker Projection list has more to do with the weakness of the clear first-round picks available in this draft than with his own merits. Strong's projection is nearly exactly average -- the average Playmaker Score for a first-round wide receiver prospect is 80.5 percent, only two ticks below Strong's score.

Overall, Strong's production was good but not great, given that the Arizona State Sun Devils passed the ball a healthy 467 times. Strong had a nice combine, but that only helps his projection a little.

Sammie Coates, Auburn

Playmaker Projection: 507 Yards
Playmaker Score: 89.9%

Sammie Coates' Playmaker numbers are favorable. However, there are also ample reasons to be skeptical of his prospects.

When Coates was a sophomore, 2.5 percent of Auburn's pass attempts were touchdown passes to Coates. Coates similarly dominated his team's passing attack as a junior, gaining 3.16 yards per team attempt. Coates also averaged a superb 21.4 yards per catch and is entering the draft as an underclassman.

However, Playmaker may be overstating Coates' prospects because, while his rate stats were good, his total stats were not that impressive -- only 741 receiving yards and four touchdowns as a junior. Auburn averaged less than 300 passing attempts per year while Coates was on the roster. Indeed, Coates' numbers may simply be a function of an offense where a few deep passes to Coates served as a change-of-pace, rather than a true reflection of Coates' ability as a wide receiver. In that regard, Coates is hauntingly similar to Stephen Hill, a huge bust that Playmaker loved. Coates' scouting report is similar: like Hill, Coates is a limited route-runner with inconsistent hands.

Not surprisingly, there are disagreements in the draftnik community about where Coates will be chosen. These projections are based on NFLDraftScout.com's forecast that Coates will go in the late first or early second round, but Scouts Inc. projects Coates as an early third-rounder.

Breshad Perriman, Central Florida

Playmaker Projection: 491 Yards
Playmaker Score: 87.7%

Breshad Perriman, the son of former Detroit Lions wideout Brett Perriman, was a consistently productive receiver for the Central Florida Knights during his career, with a stellar junior year that included a huge drop-off at quarterback from Blake Bortles to Justin Holman. Perriman enters the draft as an underclassman, has an impressive 19.5 yards per catch average, and his other Playmaker numbers are solid.

Nelson Agholor, USC

Playmaker Projection: 454 Yards
Playmaker Score: 88.6%

USC has been a bit snakebitten when it comes to sending wide receivers to the NFL, producing first-round busts such as Mike Williams and R. Jay Soward. Indeed, since sending Keyshawn Johnson to the NFL in 1996, USC has seen 16 of its wide receivers drafted, and its most successful pass catcher was one-hit-wonder Steve Smith (although the jury is still out on Marqise Lee and possibly Robert Woods).

That said, while Agholor is far from a sure-fire prospect, there is also reason to believe that he could be USC's most successful product in years. Agholor was a reasonably productive receiver in college, with 1,313 receiving yards, 12 receiving touchdowns, and four rushing attempts. Although it may seem silly, history has demonstrated conclusively that wide receiver prospects who are athletic enough to at least garner a few opportunities to contribute on the ground are more likely to succeed at the NFL level. Moreover, while still talented, USC's roster is no longer filled top-to-bottom with all-world prospects as a result of the NCAA's sanctions. Therefore, to the extent USC's prior receivers seemed better than they were because they were surrounded by so much talent, those problems are less likely to affect Agholor.

Dorial Green-Beckham, Missouri/Oklahoma

Playmaker Projection: 438 Yards
Playmaker Score: 67.7%

Assuming that Green-Beckham is drafted in the first round as expected, Playmaker projects him to under-perform his draft position. Although Green-Beckham's Playmaker Rating of 67.7 percent means that he is a stronger prospect than the average "drafted" wide receiver, wide receivers drafted in the first round average a Playmaker Rating of 80 percent.

Green-Beckham receives credit for entering the draft as an underclassman, as underclassmen are more likely to succeed than senior entrants even after controlling for draft position. However, Playmaker is skeptical because Green-Beckham was not particularly impressive in either of his two seasons of college football. During his best season, his sophomore year, Green-Beckham had only 883 receiving yards, even though the Missouri Tigers passed a healthy 414 times. Green-Beckham was not even the Tigers' No. 1 receiver: his teammate L'Damian Washington edged him out with 893 receiving yards. In contrast to Green-Beckham's high grade this year, Washington went undrafted in 2014, and has been since kicking around various practice squads.

None of this means that Green-Beckham is a surefire bust. Indeed, Playmaker sometimes has trouble with players like Beckham who have had limited college action, and thus, are less likely to produce a sample size sufficiently meaningful to forecast their futures. That said, given Green-Beckham's relatively low Playmaker Rating, NFL scouts should be doubly sure that Green-Beckham has a skill set that is likely to translate into NFL success.

Kevin White, West Virginia

Playmaker Projection: 406 Yards
Playmaker Score: 39.1%

Kevin White may be the first wide receiver selected in the NFL Draft, but Playmaker is down on his prospects.

The case against White begins with the fact that he is entering the NFL as a senior, rather than an underclassman. White's total numbers for his senior year were good: he had 1,447 receiving yards and 10 touchdowns. However, the numbers for the senior wide receivers who actually succeed are often much better than White's, especially considering that the West Virginia Mountaineers were a somewhat pass-happy offense with 534 attempts. Kendall Wright, for example, entered the draft as a senior, but had more impressive totals: 1,663 receiving yards in only 424 team passing attempts. Last year's Jordan Matthews (who also came out as a senior) had similar total yards numbers (1,477), but did so in fewer team passing attempts (376).

In White's defense, part of the reason that he exhausted his eligibility is because he played his first two years at junior college. Although the sample of drafted junior college wide receivers is limited, the data available suggests that the underclassman factor does not cause Playmaker to under- or overproject junior college wide receivers to any significant degree.

White may have impressed with a 4.35-second 40-yard dash, but among combine participants the 40-yard dash has historically had about as much bearing on success as a wide receiver's skill at needlepoint or pastry making. The vertical jump is actually the more important drill, and even the vertical jump's predictive value is miniscule. Although Playmaker certainly could be off on White, do not be surprised to see White's name, along with Marquise Goodwin and Jacoby Jones, in the inevitable stream of articles around the 2018 combine about players who ran fast 40-yard dashes but made little impact in the NFL.

DeVante Parker, Louisville

Playmaker Projection: 404 Yards
Playmaker Score: 45.9%

The appropriate question might not be why Parker's Playmaker is so low, but why Parker is even considered a first-round candidate in the first place. Parker is coming out as a senior and he does not have White's excuse that he was in junior college for two seasons: Parker has been playing college football for a full four years. Parker has never topped 1,000 yards once in those four years.

It's not as if he had a weak supporting cast either: as a junior he had the benefit of catching passes from first-round pick Teddy Bridgewater at his pre-NFL peak. A future star NFL wide receiver should have put up video game-like numbers in these conditions, but Parker produced only 885 yards. Although Parker put up some decent per-game numbers as a senior after returning from an early injury, prospects with a similar pattern of production have typically not fared well. The list of drafted wide receivers who played in less than ten games as a senior with an average of more than 100 yards per game includes Ryan Broyles, Eugene Baker, Will Blackwell, Kelley Washington, Peter Warrick, and Chase Lyman.

In any event, senior performance is arguably moot. If Parker had the talent level of a Julio Jones, he would already be in the NFL.

Tyler Lockett, Kansas State

Playmaker Projection: 387 Yards
Playmaker Score: 80.7%

As a contrast to Parker, Tyler Lockett is a good example of the level of college production that is necessary to overcome the historical trends that run against senior wide receivers. During his best season, Lockett averaged 4.02 receiving yards every time his team attempted a pass. That figure is nearly twice as large as Parker's. Although Parker had some impressive per-game numbers in a shortened senior year, Lockett actually sustained his numbers over the course of both his junior and senior years, which is an arguably more impressive feat.

Lockett's advantages over Parker's prospects from a purely analytical standpoint are not quite enough to trump the equally strong historical bias in favor of players with high draft projections. Lockett is projected to come off the board in either round two or three, while Parker is a near lock to go in the middle of the first round.

Top Sleepers

Titus Davis, Central Michigan

Playmaker Projection: 306 Yards
Playmaker Score: 85.9%

Davis rates highly due to his ability to score touchdowns. He scored 13 touchdowns as a senior even though the Chippewas passed fewer than 400 times, giving him a peak touchdown/attempt metric that beats all wide receivers in this draft class save Amari Cooper and Tyler Lockett. Davis also produced no shortage of big plays, averaging 18.2 yards per catch over the course of his career. And he had 15 career rushing attempts, which is fairly high for a player who was consistently rostered as a wide receiver.

The knock on Davis is an overall lack of explosiveness (further corroborated by a mediocre combine performance), which could cause problems getting open or gaining yards after the catch. The scouting report on Davis, however, stands in stark contrast to his numbers, which suggest that he is a playmaker capable of gaining large chunks of yards and scoring touchdowns.

Stefon Diggs, Maryland

Playmaker Projection: 306 Yards
Playmaker Score: 89.5%

Stefon Diggs, a five-star recruit out of high school, had a great freshman season, catching 54 passes for 848 yards and six touchdowns, while also rushing for 114 yards on 20 attempts. These numbers are more impressive in context, as Maryland only attempted 304 passes that year and Diggs was catching passes from a trio of freshman quarterbacks. Diggs is only regarded as a mid- to late-round prospect by conventional wisdom because his sophomore and junior years were marred by injury and inconsistency. However, one of Playmaker's lessons is that a wide receiver's "peak" season in college is a better indicator of success than the sum total of the wide receiver's career. Playmaker includes a penalty for Diggs' fall-off from his freshman heights, but that penalty is comparatively small to the penalty apparently imposed by more traditional evaluators, who generally see him as a mid-round afterthought.

Potentially encouraging for Diggs is that he maintained fairly strong per-game numbers in the latter two-thirds of his college career, and that his total numbers only suffered due to games missed for a broken leg and a lacerated spleen. It's true that these injury concerns could return to haunt Diggs' professional career, but at the low price of a fourth-round pick, the upside that Diggs teased as a freshman is well worth the gamble.

DeAndre Smelter, Georgia Tech

Playmaker Projection: 235 Yards
Playmaker Score: 86.4%

In this post-Stephen Hill world, it is hard to get too excited about Georgia Tech wide receivers. Hill, like Smelter, played in a unique offense which consisted of mostly running plays and a limited passing game that funnels its few passes to a singular wide receiver. This same dynamic led to Hill receiving a massive Playmaker Score, even though he was not very good. Smelter is not nearly as exciting as Hill, despite enjoying many of the same advantages, which makes it hard to read his Playmaker Score as anything other than a statistical quirk.

Below is a table with Playmaker Projection and Playmaker Rating for every qualifying wide receiver who was invited to the NFL combine:


Name College Proj.
Round
Playmaker
Projection
Playmaker
Rating
Amari Cooper Alabama 1 643 95.4%
Jaelen Strong Arizona St. 1 514 80.7%
Sammie Coates Auburn 1–2 507 89.9%
Breshad Perriman UCF 1–2 491 87.7%
Nelson Agholor Southern California 2 454 88.6%
Dorial Green-Beckham Missouri 1 438 67.7%
Kevin White West Virginia 1 406 39.1%
DeVante Parker Louisville 1 404 45.9%
Tyler Lockett Kansas St. 2–3 387 80.7%
Devin Smith Ohio St. 2 381 73.8%
Devin Funchess Michigan 2 356 71.2%
Phillip Dorsett Miami (FL) 2–3 309 61.3%
Titus Davis Central Mich. 5 306 85.9%
Stefon Diggs Maryland 5–6 306 89.5%
Rashad Greene Florida St. 2 288 38.2%
Tony Lippett Michigan St. 3–4 278 67.7%
Darren Waller Georgia Tech 4–5 235 69.7%
DeAndre Smelter Georgia Tech UDFA 235 86.4%
Chris Conley Georgia 4 207 52.1%
Justin Hardy East Carolina 3 203 28.8%
Ty Montgomery Stanford 5 190 59.6%
Kenny Bell Nebraska 4 180 40.0%
Jamison Crowder Duke 4–5 171 42.0%
Antwan Goodley Baylor 6–7 153 62.9%
Jamarcus Nelson UAB 6–7 134 56.5%
Da'ron Brown Northern Ill. UDFA 112 56.5%
Dres Anderson Utah 6 105 36.5%
Josh Harper Fresno St. 5–6 91 24.0%
Mario Alford West Virginia 7 90 40.0%
Rannell Hall UCF 6 81 25.9%
Vince Mayle Washington St. 5 80 15.2%
Devante Davis UNLV 7–UDFA 80 44.6%
Ezell Ruffin San Diego St. UDFA 51 31.9%
Davaris Daniels Notre Dame 7–UDFA 50 26.6%
Geremy Davis Connecticut 7–UDFA 26 17.8%
Deon Long Maryland 7–UDFA 24 18.2%
Chris Jones Alabama UDFA 0 4.2%
DeAndrew White Alabama UDFA 0 9.5%
Cam Worthy East Carolina UDFA 0 15.8%
Keith Mumphery Michigan St. UDFA 0 11.2%
Kaelin Clay Utah UDFA 0 9.5%

Portions of this article appeared previously on ESPN Insider. Playmaker projections may differ slightly from those on ESPN Insider because these numbers use NFLDraftScout.com projections of draft position rather than Scouts Inc. projections.

Posted by: Nathan Forster on 02 Apr 2015

15 comments, Last at 14 Apr 2015, 1:50pm by Matthewaz28

Comments

1
by bubqr :: Thu, 04/02/2015 - 12:36pm

The Playmaker score will definitely face a real test this year considering both Kevin White, DGB and Devante Parker are expected to go much higher than expected by the model.

It would be interesting to indicate to which player each prospect is most comparable to, using both playmaker scores/ratings + measurements and combine results if possible.
On that topic, I can’t recommend enough http://mockdraftable.com/players/2015/ - Those spider graphs are a great way to have an overview of a player’s physical profile (at least for people with a visual memory like mine).

2
by Aaron Schatz :: Thu, 04/02/2015 - 12:58pm

The ESPN version of the article had such similarities though it is hard to get players who are truly similar in every way.

For the three players you mentioned and Cooper:

Cooper: Marvin Harrison, A.J. Green
White: Michael Floyd, Donnie Avery
Parker: Bryant Johnson, Peter Warrick
DGB: Jonathan Baldwin, Yatil Green

3
by Dan :: Thu, 04/02/2015 - 2:21pm

How does Playmaker account for missed games?

DeVante Parker had 4.32 yards per team attempt this year in the 6 games that he played (which would be the highest number in this draft class), but he'd drop to a mediocre 2.00 yards per team attempt if you include the 7 games he missed.

5
by Nathan Forster :: Thu, 04/02/2015 - 8:56pm

It does not account for games missed, and this is purposeful.

I actually looked at this issue during the reboot. When I started I thought that the best way to evaluate wide receivers would be to scale the yards-TDs/team attempt by games played. So, for example, where a prospect's team passed 400 times, and the wide receiver only played in half of his team's games (like Parker), the attempts were reduced to 200. However, it turned out that ignoring these partial seasons made the model better.

The best reason I can come up for this is that Playmaker's use of the "peak" season eliminates many of the problems that this would cause. Dez Bryant is a good example. He was dominant in the three games he played as a junior, but obviously had low absolute numbers because he missed all of those games. However, Bryant still would have had a strong Playmaker score because he was also good as a sophomore. So, the lack of scaling really only downgrades a specific type of wide receiver: the player who has one good "partial" season, but did not dazzle during the rest of his career. Playmaker is willing to forgive past lack of production, but demands at least a full season worth of outstanding production.

That said, there are only a few players where scaling or not scaling makes a difference, and the results are somewhat counter-intuitive, so who knows. If Parker really booms, we'll have to take another look at using scaling for partial seasons.

Another strike against Parker, in my mind, is that Parker played 36 games, had a decently-sized role in the offense, and was only a ho-hum college wide receiver. I would be more skeptical of Playmaker if Parker had missed his first three years due to junior college, injury, or if he had only had a minimal role in the offense and did not have an opportunity to shine (see Marshall, Brandon). However, for Parker, that was not the case. The team that selects Parker is betting that, even with ample opportunity, Parker was just okay for three years and suddenly bloomed as a senior and would have maintained that pace throughout his senior year (FYI, even if you prorate his numbers, Parker is still not fantastic for a first-round prospect, just average). It is equally likely that Parker just happened to have six good games in a row.

-----------
Sorry JPP!

7
by Dan :: Thu, 04/02/2015 - 10:54pm

Another question, about both the metric & Parker: have you looked into including yards per target?

Yards per target seems to be predictive of both draft position (first rounders have significantly higher YPT than later picks) and success given draft position (first rounders below 9.0 YPT in their best season from the past 6 draft classes: Michael Floyd, Darrius Heyward-Bey, Jon Baldwin, and AJ Jenkins).

Parker had an excellent 11.1 yards per target in 2013, which makes him look less ho hum. This year he got a higher share of the targets (in the 6 games he played), and still managed to increase his YPT to 12.9.

For comparison, this year Amari Cooper had 10.8 YPT, Strong 7.7, Coates 10.2, Perriman 10.5, and Agholor 9.8. (Some of these numbers may be missing a bowl game, since I got them from a post that Bill Connelly made here in December).

8
by Nathan Forster :: Fri, 04/03/2015 - 8:35am

I would look into yards/target, but to my knowledge there is no way to back test it because target data is not available for historical prospects. If you know of some source that I don't, it would be something we could take a look at.

-----------
Sorry JPP!

10
by Dan :: Fri, 04/03/2015 - 3:33pm

I've seen target data going back to the 2005 college season, which would cover about 40% of your historical sample.

4
by Anger...rising :: Thu, 04/02/2015 - 7:54pm

If the goal was to identify receivers whose teams forced them the ball at the exclusion of all others, success has been achieved.

6
by tuluse :: Thu, 04/02/2015 - 9:01pm

You know most teams have a pretty good idea of who their best players are, and they try to get them the ball.

In college there is a wider range of talent than the NFL. So if you are NFL starting quality, you have about a 95% chance of being amazing in college to the point where the team should force you the ball at the exclusion of all others.

11
by Duff Soviet Union :: Fri, 04/03/2015 - 6:53pm

Last year there was a discussion about how the most incorrect projections often belong to prospects who share the field with another NFL caliber receiver. For instance, Reggie Wayne and Santana Moss both had bad projections because they played together, and with Jeremy Shockey as well.

Agree that that's not really an issue for most college receivers.

9
by gmoney_714 :: Fri, 04/03/2015 - 11:25am

Rashad Greene is going to play way above his Playmaker rating of 38%. The kid has too many positive intangibles that can't be factored into this scoring.

12
by Duff Soviet Union :: Fri, 04/03/2015 - 6:54pm

"The kid has too many positive intangibles that can't be factored into this scoring."

Such as?

I mean, maybe that's true, but most of the time "intangibles" are just code speak for "I know this guy sucks statistically, but I really like him and want him to be good" (or vice versa).

14
by gmoney_714 :: Mon, 04/06/2015 - 11:16am

He runs excellent routes, has an amazing work ethic, blocks well, has a ton of NCAA game experience, is durable, and has a high "football IQ".

He owns every WR record at FSU and can be a star in the NFL.

13
by nath :: Sun, 04/05/2015 - 1:13am

"Although Playmaker certainly could be off on White, do not be surprised to see White's name, along with Marquise Goodwin and Jacoby Jones, in the inevitable stream of articles around the 2018 combine about players who ran fast 40-yard dashes but made little impact in the NFL."

These are odd examples to use. Jones has made a useful career for himself as a deep threat and kick returner. Goodwin's lack of impact to date has more to do with being stuck on a team incapable of throwing downfield.

I also don't know why you'd compare a potential top-5 pick to guys who were day 2 and 3 picks in this fashion.

15
by Matthewaz28 :: Tue, 04/14/2015 - 1:50pm

Is there any way to see tables like this one for past seasons, like years 2008-2010, so that results of those draft classes could be compared now that we have seen a little more of those players NFL careers?