Predicting Running Back Turnover in the NFL: Should We Have Seen the LeSean McCoy Trade Coming?
"They did what!?"
In a move no one could have predicted, the powers that be in the front office finally pulled the trigger, sending their young star running back who amassed more than 1,300 total yards the previous season to a new team. Within moments reporters, athletes, and the general public would flood Twitter with their shocked reactions, many of which ridiculed and questioned the intelligence of those in the front office.
I am of course are referring to the Cleveland Browns trade of Trent Richardson to the Indianapolis Colts.
While seemingly bizarre when it first occurred, time and the rapid decline of Trent Richardson have vindicated the Browns' bold decision. Now, less than two years later we find ourselves again staring at what, at the moment, looks to be an equally bizarre move: the trade of Philadelphia Eagles Pro Bowl running back LeSean McCoy for a player coming back from ACL surgery, Buffalo Bills linebacker Kiko Alonso.
The move blindsided much of the NFL world when it was first announced this past Tuesday. And much like the Trent Richardson deal, the trade has left many wondering: "Who could have seen this coming?"
Can We Predict Running Back Turnover in the NFL?
The NFL is a passing league. And in this league, the job of the NFL running back is no longer a glamorous one. The emphasis on passing has generated an unwillingness of NFL general managers to pay top dollar to acquire veteran talent at the position.
I'm a firm believer that in the high stakes world of the NFL, nothing happens by chance. I believe that the decisions made by the movers and shakers of the league in regards to choosing draft picks, negotiating contracts, and making personnel decisions (unless you're the Washington football team, of course) are based on a foundation of strong analytics, forecasting, and financial prudence.
So with that being said, how could we have seen these blockbuster trades happening? Was there truly a rational method to the madness? To answer this I began by asking the question: "What factors influence running back turnover in the NFL?"
For the purposes of this article, I define running back turnover as the event when a starting running back loses his starting job to another running back due to performance (i.e., non-injury or suspension) related issues. Going back to the 2013 and 2014 seasons there were a total of nine and 13 such instances, respectively.
From this I made an intuitive hypothesis: that turnover was likely to occur where the incumbent running back had poor rushing productivity the prior season relative to the run blocking abilities of that player's offensive line.
Testing the Hypothesis
To test this hypothesis I went back to the 2013 and 2014 seasons and ranked each incumbent starting running back (1st being best, 32nd being worst) based on their ability to increase their team's likelihood of putting points on the board as measured by our total Rushing Net Expected Points (NEP) metric from the previous season.
From these graphs, I then marked each running back who lost his starting job to a backup or newly acquired running back sometime during the following season.
Below are the graphs from this analysis.
What we see is that players falling in the bottom half of all starting running backs as measured by total Rushing NEP had a strong tendency to lose their starting job the following season. Indeed, eight out of the nine running backs (88.9%) who lost their job in 2013, and nine out of the 13 running backs (69.2%) who lost their job in 2014 had a Rushing NEP from the previous season that ranked them in the bottom half of the league.
When looking at the lower right quadrants -- these same "bottom-half" running backs whose offensive lines ranked in the top half of the league -- we see a very strong enrichment for running backs who were replaced as starters.
From this analysis, we see that Rushing NEP relative to offensive line run blocking grade was a strong predictive factor for running back turnover.
When viewed in the context of the analysis above, the trade of Trent Richardson now makes a bit more sense. The Browns were trading away a player who had a high likelihood of not being the starting running back for their team by the end of the season and receiving a coveted first round draft pick the next year in return.
The LeSean McCoy Trade and Forecasting Running Back Turnover for 2015
Are the factors influencing the decision to trade Trent Richardson relevant to the Lesean McCoy-for-Kiko Alonso trade? To answer this question I applied the same analysis I used above on the starting running backs at the end of the 2014 NFL season. Below are the results from this analysis.
The first shots fired of the upcoming 2015 NFL offseason regarding the fates of these running backs again echoes our findings from the previous two seasons: LeSean McCoy (traded), Steven Jackson (released), and Fred Jackson (likely replaced by the McCoy acquisition) fell into the bottom half of the league as ranked by total Rushing NEP. McCoy himself, falls into the dreaded bottom-right quadrant.
From this we get a feeling of déjà vu.
Was McCoy traded for precisely the same reasons the Browns parted with their first round draft choice early in the 2013 season? Only time will tell, but based on this analysis, the acquisition of a potential Pro Bowl linebacker, and the strength of the 2015 NFL running back draft class, including Doak Walker Award winner Melvin Gordon, the chances of the Eagles being on the winning side of the equation are pretty good.
We see that the ramifications of this analysis extend beyond just the McCoy trade. A number of names including Frank Gore, Isaiah Crowell, and Stevan Ridley fall into this quadrant alongside McCoy, which may signal an end to their reigns as NFL starting running backs (at least for their current teams) as well.
Again, only time will tell, and many other factors are in play, including which free agents are signed and which rookies are drafted this upcoming April by each team. But based on the metrics presented above, the chances for these players holding down their starting jobs next season are slim.
Many analysts tend to forecast running back production based on a number of factors, most famously basing their predictions on what side of the 30-year-old line a player falls on. From this analysis, it's tempting to suggest that perhaps it's time we start to judge these athletes on what side of the NEP line they fall on instead.