F= MA: Secret Ingredient for Exceptional Wide Receivers?
My apologies for the time it has taken me to compose the important conclusion to this series on what characteristics might make for exceptional NFL wide receivers.
In today’s pass-happy league, these are extremely important positions on one’s roster.
We continue with our analysis of 495 NFL wide receivers from 1999-2012 as composed by Dr. Peter-Lawrence Smith and kindly shared by Armando Salguero from the Miami Herald.
 |
| Dolphins Wide Receiver Jarvis Landry: Wilfredo Lee/AP |
That data includes both the pre-career Scouting Combine or Pro-Day performance of each man and their eventual performance in the NFL.
We covered many of the details in the preceding column. We also concluded in the last column, based on the numbers, that evaluating Dr. Smith’s designation of elite wide receivers that we could find no definitive correlation of those players with their 40 yard dash times from the combine. This flies in the face of conventional wisdom which holds that speed is extremely important.
That’s one reason that Sammy Watkins was so sought-after in this year’s draft (he’s very fast at 4.34 seconds). That’s also a reason why receiver Jarvis Landry, which such reliable hands and good production for LSU fell to a point where the Dolphins could draft him. He was timed at a pedestrian 4.77 seconds in the 40 yard dash (However, in fairness, it is important to note that Landry ran it in 4.58 seconds at LSU’s pro day after his injured hamstring had healed).
Does that doom Landry in the pros? And how important is speed to NFL production form a wide receiver anyway? If speed was everything, then wouldn’t all elite wide receivers be track stars?
First, let’s start with the premise that the single most important measure of production from a wide receiver is touch downs per game played. Sure, yards per game is important as are catches, yards per reception and other characteristics. While there are many metrics, let’s assume that touchdowns per game are important.
To eliminate bias from receivers that were never really played, let’s further limit our field of assessment to players that were in more than 16 games in their career– the equivalent of a full season of play. That immediately removes the newest players with a limited track record and drops the numbers of receivers in our assessment from 495 to 261 receivers.
We’ll then use Dr. Smith’s data to find out how important the various combine measures might be to eventual NFL performance. We include the 3-cone drill as well as shuttle times to see if we these agility measures might be important . We convert weight and 40 yard dash times into mass (kg) and speed (meters per second) for reasons that will become apparent shortly.
• height ( inches)
• vertical leap ( inches)
• broadjump (inches)
• shuttle time (seconds)
• cone time (seconds)
• mass (kg)
• speed (m/sed)
Requiring that all our candidates have all these data further reduces the sample down to 175 receivers– still a healthy sample size.
We then subjected each of the above characteristics to an Analysis of Variance (ANOVA) to see which of the above variables seemed significant to touchdown productivity per game. There were only three of the variables that were statistically significant at a 90% level. Vertical leap, broadjump, cone and shuttle times were not significant. In order of significance, they were:
• Mass (weight in kg)
• Height (inches)
• Speed (m/sec)
However, putting these into a multiple-regression revealed that height in inches was not significant once mass was included– the two characteristics were related and mass was much more important of the two. However, even with mass included, speed showed up as being very important– so important that both together explained about 7% of the touchdown production of NFL receivers. That may not sound like much, but remember we are trying to find a way to understand what might give an edge in evaluating talent at the combine. What did it say?
Look for big, fast receivers.
Realizing that mass and speed were both important, I suddenly remembered high school physics.
Force = mass times acceleration
And acceleration was speed squared.
Could it be that the statistics were just finding out what Newton had concluded centuries before: that NFL receivers that would be most difficult to stop from reaching the end-zone would be those who were heaviest and fastest?
Thus, I created a new variable which was Force:
Force = mass (kg) * m/sec^2
Where:
Force is in Newtons (appropriately enough)
The regression analysis then revealed that the strongest (and only variable) with statistical strength to explain the long term success of NFL received in achieving touchdowns each game was Newton’s Force. Once Force was included in the regression, neither mass nor speed was important as single added variables.
Force was the single and most powerful explanation for touchdown efficiency and not speed alone.
Moreover, if I repeated the regression with Dr. Smith’s elite players
entered as a dummy (0,1) variable, the elite designation immediately shows as important,
but Force is still statistically significant.
Finally, if we regress on only the 18 elite players, we still find that Force remains significant. Indeed, it explains about 25% of the variation in touch down productivity among Dr. Smith's elite receivers.
I then plotted the receivers graphically against their efficiency (TDg) versus their indicated Newtons from the combine data. These data are shown below:
If studied carefully, the plot can be quite interesting. It shows a phenomenal amount of scatter as well as a very general trend that is captured by the statistical model. (The red line is the predicted change in TD efficiency with greater force from the regression). If there was perfect correspondence between force and TD efficiency the data would neatly fit across the diagonal of the plot– it doesn’t..
The plot shows that Calvin Johnson’s (0.62 TDg , 7664 Newtons) success fits the model very well, the same way it also predicts that Craig Yeast (0 TDg and 4981 Newtons) would fare poorly. It also does a nice job showing how Julio Jones, Andre Johnson and Vincent Jackson may do well because of physical attributes. Mike Wallace also looks like a good prospect from the chart.
But the plot also hints at much more important relationships from what it doesn’t show. As Malcomb Gladwell so famously indicated in
Outliers, those players that both over and underachieve on the plot are even more interesting than those that fit the model. Why is that?
Remember the combine performance characteristics were only able to find to two characteristics, speed and mass that mattered. And they only explained 7% of the variation. By studying those player who both over and under achieve we may get insight into what matters that we can’t measure in Scouting Combine statistics.
Taking a look, the underachievers include Legedu Naanee who played for the Chargers, and Carolina before signing a one year deal with the Miami Dolphins on April 17, 2012, He was cut on October 2, 2012 after posting only one catch for 19 yards in his four-game tenure with the team. Why did Naanee do so poorly? One commentator: “Hands of stone...” Another underachiever: Greg Little who is now with the Raiders after being cut by the Browns. Reasons for poor performance: a history of inconsistency and drops.
 |
| Legedu Naanee was a bust with the Dolphins (Ronald Martinez/Getty) |
Delanie Walker and Niles Paul appear to be underachievers, but their low performance likely stems from transition from wide receiver to tight end. A third underachiever is Darrius Heyward-Bey who is now being given another shot with the Pittsburgh Steelers. Heyward-Bey’s reputation? Bad hands. So we are immediately getting a picture of the ingredients for under-achieving: dropping passes.
What about the over achievers?
Most obvious of these are A.J. Green, Greg Jennings and Desean Jackson. What do they have in common?
A.J. Green (Bengals): great hands and a knack for catching the ball in stride
Greg Jennings (Vikings): great hand skills and route running
DeSean Jackson(Washington Redskins): great hands, deep ball speed, lateral quickness
So, what do all three over-achievers all have in common:
great hand skills. And that is the inverse of those who underachieve. Poor hand skills and a lot of drops turn you into a physically talented bust.
 |
| Greg Jennings with the Vikings is known for his excellent hands |
So, an obvious, but important conclusion. Speed and body mass matter a lot, but great hand skills appear to be very important as well, but difficult to assess in a non-NFL evaluation environment.
 |
| Sir Isaac Newton would, no doubt, be pleased to learn his 2nd second law of motion governed NFL receiver potential |
|
So, to cap off this segment, what does this evaluation say about Jarvis Landry and the Dolphins?
First let’s assume that his true speed is that from his uninjured LSU pro-day: 4.58 seconds for 40 yards or a speed of 7.99 m/sec. His weight is 205 lbs (93 kg). This creates a Force level of about 5937 Newtons. This is similar to Greg Jennings (8.3 m/sec and 89 kg, 6119 N) or even more to the indefatigable Reggie Wayne (8.2 m/sec, 90 kg; 6067 N).
 |
| Jarvis Landry as seen in recent Organized Team Activities (Miami Dolphins) |
Jarvis Landry was famous at LSU for his hands skills and circus catches. My conclusion is that Landry still may be an outstanding NFL receiver, regardless of his relatively slow combine time on the 40 yard dash. The biggest question: will he be able to achieve a degree of separation against the faster cornerbacks in the NFL to use those hand skills.
That will be important to evaluate in preseason. We shall soon see.
