December 7, 2016

Testing Bill Simmons’ Theory on Running QB’s

By Bill Barnwell, Football Outsiders – special to BSMW Patriots Game Day

Resuming normal service this week, I’m going to take a look at something Bill Simmons brought up in his column about the struggles of Daunte Culpepper, struggles which led to Culpepper’s benching last week before his game against the Patriots’. In the column, Simmons writes the following:

Running QBs are like professional wrestlers and porn stars. In other words, it’s such a taxing profession on so many levels, and you end up taking such a pounding, there’s only a five- or six-year shelf life before things turn sour.

Lamentations of Gino Hernandez and Traci Lords aside, I thought about the idea for a minute, and it didn’t seem too absurd. I felt even more agreeable to the idea after Simmons provided actual data – granted, cherry-picked data, but that’s still eight or nine years ahead of Joe Theismann – in the defense of his argument.

Now, I’m aware that the internet backlash to our dear friend Mr. Simmons is mighty strong right about now, particularly after a rather desultory comment towards the sports blogosphere in a recent chat. Now, Bill Simmons is about eight thousand times the writer I am – he can say whatever he wants. That being said, isn’t that a little like NFL announcers insulting fantasy football players? It just seems strange to insult your core constituency. I mean, you don’t see Dennis Hastert saying people who go to church on Sunday are dorks, right? I digress.

Interested in its validity, I decided to do a more rigorous test of Simmons’ theory and what it might mean for the career paths of quarterbacks. I had to put several limitations on the quarterback pool to make sure we were working with the quarterbacks Simmons is talking about, which meant I filtered the sample down to quarterbacks had both:

  • Started their career during or after 1978 (the advent of the sixteen-game schedule), and
  • Thrown more than 1500 career attempts (since we’re not comparing Danny Kanell to Dameyune Craig here, but instead legitimate starting quarterbacks and their career paths, which requires several years of data to work with)

That brought the pool down to 83 different quarterbacks. I toyed with different ways of delineating the quarterbacks into mobile and immobile groups, not wanting to just separate the two willy-nilly. What I found worked best was using the number of rush attempts per game a quarterback had on a season-by-season basis, and then separating players from the mass as such.

I defined a “rushing” quarterback as one who averaged more than 4.5 carries per game for any two seasons in his career. That yielded a list of quarterbacks I was pretty comfortable with:

rushingQB.jpg

Not particularly coincidentally, this includes each of the players that Simmons provides as examples of the running quarterback problem in his column. It’s also worth noting that Michael Vick doesn’t appear in this group because he falls 67 attempts short of the 1500 pass qualifier. If he’d made it past the bouncer, he would’ve easily qualified for this group.

The flat-footed quarterbacks, on the other hand, were even harder to define; after all, it’s very easy for a quarterback to rack up very few carries per game by simply not playing — alternately, it’s easy for a quarterback to get three or so carries a game solely by downing the ball as a backup. With that in mind, I made the qualifications slightly harder for this group. The quarterbacks listed below had five seasons where they averaged below 1.75 carries per game, throwing at least 150 attempts in each. That yielded a group of, well, guys with creaky knees.

tortoiseqb.jpg

I am sure there are arguments you can make for putting a particular quarterback not listed here in one group or the other, but I think these are two pretty representative groups of quarterback.

After separating the groups, I calculated the performances of each quarterback in the respective season of his career – that is to say, I calculated how each quarterback did in the first season of his particular career, added that to the performance of every other quarterback in their particular career, and then produced an average result for all the quarterbacks in their first seasons. I did that for every group of seasons available, stretching all the way to 19 (Dave Krieg and Vinny Testaverde, before you ask) before I was done. I calculated this for all quarterbacks (incorporating all 83 quarterbacks), the running quarterbacks (listed as “Fast” in the charts below), and the slow ones (listed, as you might expect, as “Slow”).

I’ll include all the results in a table at the end of the article, but I chose to focus on three aspects of performance: yards per attempt, QB rating, and since the original article was about Culpepper’s fantasy dropoff, fantasy points.

chart.jpg

As you can see, the performance of the faster quarterbacks cut out after season 15, since none of the rushing quarterbacks in the study have made it that far. This may be because seven of the eleven quarterbacks in the group are still active, and haven’t had a chance to make it to their fifteenth season as of yet. Regardless, what the data shows is that the quarterback performances are actually almost the opposite of what Simmons mentions in his column: the running quarterbacks underperform both the average quarterback and the slower quarterback until their sixth season, at which point they spend several outperforming the quarterbacks until a big drop around Season 11 (which includes a mediocre Steve Young season, a mediocre Jeff Blake season, Mark Brunell’s first aborted season with the Redskins, Randall Cunningham’s final half-year as an Eagle, and Steve McNair’s last year with the Titans. In other words, Donovan McNabb is in trouble when 2009 rolls around). You could chalk this up to a number of things — maybe the running quarterbacks are adapting to an NFL system, or becoming “pocket” passers, but I wasn’t expecting that data trend whatsoever.

QB Rating shows a variation on the story.

chart2.jpg

This time, the faster quarterbacks have a better showing, but they still don’t branch out from the pack of other quarterbacks until Seasons 6-10 roll around. Could it be that speedier quarterbacks elude sacks and throw fewer interceptions than slowpokes? The following chart shows the average number of attempts it takes a quarterback in each of these groups to throw an interception, followed by each group’s cumulative average:

attempts.jpg

Pretty obvious that rushing quarterbacks do throw fewer interceptions than pocket passers — a relatively hidden advantage of their performance up to this point.

Finally, to answer Simmons’ question, do these stud quarterbacks actually perform worse after a few years in fantasy points?

chart3.jpg

The short answer is, well, it’s debatable. The running quarterbacks don’t really seem to fall off a cliff in five or six years, like Simmons says, but he’s not far off. Somewhere around eight to nine seasons appear to be the limit for rushing quarterbacks — that big jump in the last year you’ve been seeing is the confluence of Mark Brunell’s revitalization last season with the Redskins and Randall Cunningham’s gigantic 15-win season with the Vikings. What does appear to be true, though, is that rushing quarterbacks are peaking very early into their fantasy careers, while slower quarterbacks do so slightly afterwards. It’s interesting that they both see the same dropoff after eight or nine seasons, though.

I’m inclined to say that Simmons’ hypothesis, for fantasy purposes, is pretty accurate. Rushing quarterbacks peak earlier into their career than I think a lot of people, myself included, realize: a sobering thought for those with Michael Vick, and a happy one for those who have Vince Young in a keeper league.

And now, a data dump.

datadump.jpg

Leave a Reply