There is a new analytics department with GM Dennis Hickey and the Miami Dolphins. No longer a matter of speculation, the rumored “crystal ball” group’s existence is verified. Hickey himself admits as much:
http://www.miamiherald.com/2014/04/16/4060760/technology-embraced-by-miami-dolphins.html
While we’re unlikely to learn most anything about the Dolphins equivalent to Black Ops, we can imagine there are some pretty sophisticated number crunching gurus shredding numbers behind the curtain.
Recently, intrepid Miami Herald sports columnist, Armando Salguero, posted intriguing data from Dr. Peter Lawrence Smith on wide receiver prospects for the Miami Dolphins. Dr. Smith is a Postdoctoral Research Fellow for Clinical Sciences at St. Georges University of London. And like me– obviously possesses more than a passing interest in football– and numbers!
Dr. Smith’s data summarize the performance of wide receivers in the NFL from 1999 - 2012 along with data on their post college combine scores at the various competitions. Armando’s post of Dr. Smith’s provocative findings was an indirect challenge to the Dolphin’s analytics team:
With such an approach, how did the Dolphins come to select wide receiver Jarvis Landry in the second round of the draft?
Sure, Landry was known as an extremely tough and sure-handed competitor at LSU, running excellent routes, with a willingness to take hits and fight for yardage. His big college reputation: he seldom drops passes. Moreover, Landry displayed excellent blocking skills similar to Anquan Boldin. ESPN’s Todd McShay, praised the Dolphin’s 2nd round pick and believes Landry could be one of the best draft values.
 |
| Jarvis Landry fights for the ball: (photo: Bill Feig: The Advocate) |
Landry is not an athletic super star.
In particular, his 40 yard dash time in the Indianapolis combine was 4.77 seconds– one of the slowest. At 5 feet, 11 inches and 205 pounds, he is neither exceptionally tall nor heavy. Later, on his pro day Landry ran a 4.61 forty yard dash twice. Although better, that’s none too fast. Dr. Smith expressed questions about Landry’s speed in his analysis of his data and what that would mean for his potential in the NFL.
His evaluation found that while “slow receivers” like Landry sometimes did well in the SEC colleges, that potential did not often carry over to the NFL. Davone Bess had a similarly slow speed (4.64) and was lighter ( while being able to be productive in the NFL for several seasons). But...
"Landry could be the slowest second-round pick by as much as 0.17 seconds,” Dr. Smith commented."Landry's 4.77 combine time is slower than any of the 284 receivers in the data set.”
Further, to Smith, that statistic does not bode well for Landry "According to my analysis Landry is unlikely to have success and drafting him in the second round is a reach." You can read more of Dr. Smith’s logic here:
http://bit.ly/1k1646K
Salguero kindly made Dr. Smith’s Excel data file available, which I promptly downloaded and moved into STATA– a statistics package I frequently use in my work for the Department of Energy at the University of Central Florida.
While I was sure Dr. Smith would have looked at the data in detail, I wanted to do my own examination and see what I could find out. I work with statistics very often in trying to ferret out useful information from messy data and I am familiar with many of the tools. For such purposes, I often use step wise multiple regression coupled with Analysis of Covariance (ANCOVA) or other non-parametric procedures to sort out complicated and often, interacting relationships.
Was speed as important as Dr. Smith assumed it to be?
And what would the full data on the 495 wide receivers in the NFL from 1999 - 2012 tells us about what performance measures at the combine might translate to success on the grid-iron?
I was surprised from what I learned and I think sports fans might be too...
Hint? Examine the title of this column.
Until next time.
No comments:
Post a Comment