Let's talk about the Cowboys slow starts
Throughout the 2019 season, I’ve found myself wondering how much to make of the Cowboys pattern of routinely falling to big early deficits.
In their eight losses this year, the Cowboys have scored a total of 17 first-quarter points—just over 2 points on average in the first quarter of each losing effort. Meanwhile, their defense has given up 52 points in the first quarter of these same games.
These early deficits are “manageable,” in the sense that trailing teams come back from a first-quarter deficit about a quarter of the time. But once a team falls behind, each additional misstep on either side of the ball can place hopes for a comeback even further out of reach—as the Cowboys know all too well.
These slow starts are all the more frustrating because, in view of their entire body of work, we can be fairly confident the Cowboys are a good football team with a very good offense.
The Cowboys have outscored their opponents by a margin of +82 points over the course of the season, a number more consistent with a 10-5 team than their current 7-8 record. And thanks to the advent of public expected-points models, we know the Cowboys pass offense has been doing most of the heavy lifting to drive that scoring advantage.
So why has a team with such good offensive results and average defensive results overall performed so much worse at the start of games?
Some plausible explanations:
Insufficient play scripts—Teams often prepare a “script” of offensive plays that is custom-tailored to attack specific weaknesses they’ve seen in opposing defenses. Could the Cowboys plan itself be dooming them to fail?
Worse quarterback play—There’s no denying that the Cowboys have gotten worse outcomes when they’ve dropped back to pass in the opening possessions of a number of their games. Is Dak Prescott just playing worse in these situations, or are there other reasons for the gap?
Sequencing luck—Football is a cruel sport where the unpredictable bounce of a goofy-looking oblong ball can determine the outcome of an entire game. But one of the main ways teams’ records can fall out of step with their underlying performance lies in the timing of their most valuable possessions. Could the Cowboys first-quarter struggles be a function of not getting timely breaks that most teams depend on?
Fortunately for us, Pro Football Focus data scientist Eric Eager recently explored some of these questions from a much more comprehensive point of view.
One of the great benefits of using data in sports is we can take common, team-level narratives and see if they hold up at the league-wide level. The bigger the sample, the more confident we can be in our conclusions. This is where we tend to want to start—trying to wrap our heads around the league-wide dynamics—before we drill down into the specifics for each team.
Generally, Eager found that a team’s performance in the more scripted portion of the game—the first 15 plays on offense or defense—wasn’t all that reliable.
In other words, teams like the Cowboys, who have been consistently worse at the beginning of games than their season-long averages suggest they should be, are not particularly more likely to see those early struggles continue.
From Eager’s article:
The variable importances for overall offensive and defensive EPA were more than 10 times that of either during scripted plays. Hence, if you want to wager on first-quarter totals, just ask yourself how good the two teams are offensively and defensively (when it comes to EPA-related variables), weighing your prediction heavily towards offense.
Interestingly, while Eager found past performance in the opening of games isn’t particularly useful for predicting future early-game results, the uniqueness of an offensive scheme may be a significant predictor of first-quarter performance.
What does this mean for the Cowboys?
Overall, I think any time you see an outlier as big as the 2019 Cowboys, there are likely to be multiple explanations.
The PFF study suggests that, in most cases, teams’ first-quarter performance is extremely volatile, and it shouldn’t surprise us that a team like the Cowboys could largely be victims of dumb luck.
But Eager’s secondary finding—that qualities unique to a particular offensive scheme can predict first-quarter results to some degree—leaves open the possibility that the Cowboys gameplans have been insufficient as well.
Unfortunately, we probably won’t ever know with confidence how much these early deficits can be blamed on the gameplan or variance in quarterback play vs. simple dumb luck.
But for my part, I think we can be pretty sure that if the Cowboys were allowed to play their 2019 games over again—for better or worse—the journey would look almost nothing like this big dumb season.