Trying to predict turnovers in the NFL is a risky proposition


The NFL offseason is a time for the talking heads and the analysts to rehash some age-old and accepted talking points in previewing the season ahead. One of those popular points is that turnovers are unsustainable for a defense. They say always that you are surely bound to see a sharp regression in your turnover count following a really high number of forced turnovers the previous year. The extremes don’t last, everything always reverts to the middle, fundamentals matter most, etc.

Well, I decided to put that theory to the numbers test. But not just for regression, progression too. I looked at the sustainability of both extremes for the time period from 2008-2013, which provided me with the last five years of change data.

turnover regress_progress

Immediately, you can see that regression is not as clear-cut as most people would lead you to believe. Following a top 5 turnover ranking, 11 out of 25 teams actually still stayed in the top 10 the next year, a not-insignificant amount, almost half in fact. Only 5 teams actually even fell to the bottom 10. And certainly an 11 to 14 split is nothing conclusive.

Progression meanwhile is a little different; there’s a clear trend. Only 8 out of 25 teams who ranked in the bottom 5 in turnovers stayed in the bottom 10 the next year. 17 teams managed to improve and got themselves out of the cellar. 4 of those teams actually managed to leap into the top 10. There’s certainly more evidence of elasticity in progression, at least on the surface.

Why would progression be more clearly elastic than regression? Well, often times, teams with low turnover numbers on defense may not be all that good in other areas of defense as well. You don’t often see good defenses with abysmal turnover rankings. This could lead to an influx of new talent in the offseason or coaching changes, which logically are made to improve those numbers. On the other hand, high turnover numbers can either be a byproduct of a really good defense or can mask other deficiencies, leading to a more scattered distribution when it comes to regression data.

Now what about the sustainability of sharp changes? 3 teams went from top 5 to bottom 5 in turnovers- the 2009 Browns, 2009 Dolphins, and 2011 Steelers. The 2010 Browns rose back up to 14th, the 2010 Dolphins got worse (at 29th), and the 2012 Steelers stayed in the bottom 10 (at 25th). 3 teams also followed up a bottom 5 year with a top 5 year- the 2009 49ers, 2013 Chiefs, and 2013 Eagles. The 2010 49ers dropped all the way to 25th, and well, only this next season can tell us how the 2014 Chiefs and Eagles are going to fare. The sample size is too small and the results are too scattershot to determine anything of value here.

All of this is to say, take those analysts’ words with a grain of salt. Regression is hard to predict. Turnovers are fickle. It’s more useful to go in depth and break down the team’s roster and offseason moves. Though I must say, I wish someone would pay me to get on TV and say something like “The Patriots are bound to decline in the turnovers category next season” without any real evidence.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s