NFL BY THE NUMBERS
Toxic Differential
We have entered the age of the sports nerd when it comes to stats. No longer can we identify trends and predict success (usually after the fact) from mere mortal stats like yards allowed, average yards per pass, etc. Now we seem to have a cottage industry that seeks to create and legitimize ever more nuanced
number breakdowns. Blame it on Nate Silver.
I am a CPA by trade -- I love numbers and I love stats and I love trying to make sense of them. But let's remember -- sometimes we only see in numbers what we want to see in them.
I am a CPA by trade -- I love numbers and I love stats and I love trying to make sense of them. But let's remember -- sometimes we only see in numbers what we want to see in them.
A word of caution as I introduce you to the Toxic Differential.
Brian Billack wrote a column today for Fox Sports that explains the Toxic Differential. It is actually a pretty simple statistical analysis that combines two differentials: Turnover margin and Big Play margin. For this purpose a Big Play is a play that gains 20 or more yards.
As an example consider the following data points for Mythical Team A:
As an example consider the following data points for Mythical Team A:
Turnovers - 12
Turnovers forced- 9
Big Plays - 40
Big Plays allowed 31
Team A therefore has a Toxic differential of 6 (-3 on turnovers and +9 on Big plays).
An interesting stat --- but how well does it predict success on the field? Niners, Ravens and Broncos are at the top of the list, that makes sense. The Chiefs and Eagles are at the bottom of the list - that makes sense. (See chart below excerpted from Billack's article) But....
An interesting stat --- but how well does it predict success on the field? Niners, Ravens and Broncos are at the top of the list, that makes sense. The Chiefs and Eagles are at the bottom of the list - that makes sense. (See chart below excerpted from Billack's article) But....
How much predictive value of success is there to a stat where the Panthers, Browns, Lions and Redskins comprise half of the top eight (combined record 15-29)? What should I make of the fact that the only two 10-1 teams in the NFL do not even crack the top ten (Houston is # 11, Atlanta #14)?
Coach Billack tries to draw some pained lessons from these stats - but it rings kind of hollow. Simply put, Carolina's 3-8 record does not appear to jive with its fourth rated Toxic Differential. This is a fun stat -- but so is picking winners by which jerseys you like best.
The best stats are sometimes the most simple and require the fewest qualifications. Look back at my November 8, 2012 post on sacks. Teams that win the Super Bowl average 23% more sacks than the league as a whole. Simple. Predictive. Toxic Differential -- it tells me I should be expecting a Panthers/Browns throwdown in February before I should expect a Falcons/Texans contest. Hmmm.
No comments:
Post a Comment