One of my favorite things about football season is reading TMQ (Tuesday Morning Quarterback, for those unfamiliar) on ESPN.com. I’m not entirely sure why, because it’s gotten to the point that I skip over whole sections of his articles (namely the ‘Christmas Creep’ part; yes, I know that advertising for things happens well before it should). I think it’s because he always says some things I agree with and some things I don’t. Mostly I think I like the latter. He has a new article today, and in true TMQ tradition has already said something questionable. Namely, that running in the NFL has become ineffective, or at least not important to winning.
To be fair, TMQ mentions a couple reasons this could be. It could be due to rule changes that favor passing, or due to the spread of research that suggests that passing is simply better because of the higher potential yardage to be gained (he mentions Football Outsiders; I would prefer to mention Advanced NFL Stats). I think either or both of these is probably true. But TMQ spends time on one of his favorite stories, which is how huge offensive and defensive linemen have gotten in the past 30 years or so. He says there simply isn’t as much room to run anymore, so teams are running less.
As you might guess, I disagree. I’m using data from pro football reference (specific page here). It’s certainly true that rushing attempts have dropped recently; eyeballing a plot of team average rush attempts per game against season, attempts have dropped a fair amount since the 70’s. But TMQ’s claim that they’ve dropped because it’s harder to run implies that yards per attempt should drop as well. Here’s the plot of yards per attempt by season in NFL history, which covers 70-some years:
Now let’s agree to ignore the really old data, because so much has changed since then. Also, TMQ specifically talks about how much bigger linemen have gotten since the 70s. Here’s a graph of the past 30 years:
The raw data (open circles) isn’t as clear as in the full-history graph, but running has indeed gotten better, not worse (all of what I’m about to say is also true if you include the 70s, or move from 30 to 40 years). The solid circles are predictions from a linear model, regressing year on yards per carry. The curved diamond line contains the predictions from a quadratic model that adds year squared. The quadratic line isn’t quite significant, but regardless the squared term is positive, implying that not only is yards per carry increasing, but the rate at which it’s increasing is increasing. The important part is that it isn’t negative, which would mean that yards per carry is reaching a plateau or starting to go down; it’s only going higher. For those of you worried about multiple endpoints in choosing 30 years, I ran a correlation test (the same as a simple regression, but quicker in my stats package) between year and runs per carry, and it’s significant from 15 years (1994 to 2009) out to 40 years (1969 to 2009) in 5 year increments as well as in the whole data set, so I don’t think 30 years is special.
So what now, Mr. TMQ? It’s certainly true that linemen have gotten much larger, and it probably isn’t healthy for them in the long term. But it isn’t hurting teams’ ability to run the ball – they’ve only gotten better at it in the last 30-plus years, and they might be getting even better. It’s possible that’s due to teams adjusting to the huge guys in the middle of the field by running more to the outside. I don’t know if they ran outside more, but in 2009 at least rushes gained more yards off the end than up the middle, off guard, or off tackle (numerically at least; I took a quick look at some data from Brian Burke). So that could be the case. But all in all, I find it more likely that good teams are running less because they’re catching on to passing being more efficient.