The Tampa Bay Buccaneers ran the ball just nineteen times against the New Orleans Saints, three of those runs being Jameis Winston scrambles. Doug Martin racked up 81 rushing yards on just 11 carries, and the failure to give him the ball has quickly been identified as a major culprit in the Bucs' loss. Rick Stroud wrote about it, asking how "when the Saints ran Tim Hightower 28 times, how does the NFL's second-leading rusher only get 11 carries," while Ira Kaufman went with the statistical approach: apparently, the Bucs are 2-21 over the past five years when attempting 20 or fewer rushing plays.
The Bucs had 19 on Sunday. Does anyone think a single extra Doug Martin run would have somehow saved the day?
The fact that the Bucs run the ball less and use Doug Martin less in losses than wins is a simple consequence of the fact that they're a lot less likely to be leading and running out the clock. It's not a cause of losing, but an inevitable result. This is not a secret. In fact, it's been such an obvious truth that Football Outsiders has had it as their first bullet point in their "Basics" article for as long as I can remember. It's practically the only thing that all modern statistical NFL analysts can agree on.
And the Bucs know this. Lovie Smith does, too. When asked about it, he said that it "just happens."
"First off, how many plays, time of possession, there's a lot of things instead of just Doug didn't get the ball that many times. When you don't convert on third downs, it can stop drives. I think you have to look at the full context of it all and we end up getting the results we did."
It may not be a satisfying answer, but it's the truth. The context and game situation determine who gets the ball and how much they get it. And we can show this statistically, too. While Kaufman's statistic seems stark, the correlation's causes are obvious when you break them down by half. In the first half, when the Bucs almost always have enough time and aren't chasing a deficit, they run the ball as much in wins as they do in losses. In the second half, when context starts determining whether and who gets carries, you start seeing a split appear. The answer is clear: the Bucs run the ball less in losses, because they don't have the opportunity to run the ball in losses.
In six wins, the Buccaneers rushed the ball 92 times in the first half. In seven losses, they did so 104 times. That's barely a difference, and certainly not one big enough to be the cause of any losses. The distribution of carries is pretty even, too: Charles Sims got 25 versus Doug Martin's 60 in losses, while the ratio is 23 to 62 in wins.
In the second half, things look a little different: they ran the ball 71 times in losses in the second half, and a whopping 114 times in wins. When we look at how these numbers break down, it's easy to see why they're so skewed: they include nine kneeldowns, a further 22 carries in their dominant win over the Philadelphia Eagles, and another eight carries against the Jaguars in a game where their lead was never really in doubt. It's easy to run the ball when you're dominating -- a lot harder to do so when you're losing and need to catch up in a hurry.
Perhaps the result has to be looked at in terms of Doug Martin and Charles Sims, then? Martin got 71 carries versus Sims' 24 in the second half of wins. In losses, he's gotten 40 against Sims' 19. That's certainly skewed, but it's also easily explained by the fact that the Bucs are playing catch-up in losses, which means passing, and Sims is generally on the field for passing downs -- because he's the better back in that situation.
There's one more stat that'll show you the problems of just blandly stating the numbes and assuming they're causally related. You would expect the Buccaneers to be better at running the ball in wins than losses, right? But in losses, the Bucs have racked up 5.1 yards per carry, compared to 4.6 in wins. That's not just caused by kneeldowns: Doug Martin is 0.6 yards per carry more productive in losses than wins, and Sims is only 0.4 yards per carry less productive. This doesn't really mean much: variation happens, and obviously being more productive is better than being less productive.
It just shows you the weird things that happen when you try to seek causal relationships based on fairly limited numbers.