At the quarter mark of the NFL season, there’s enough data and film on these teams to get a decent idea of what they’ll be moving forward. Are the Dallas Cowboys actually a 1-3 level team? Or are we in for another anomaly year where they look better than their record by nearly every measure?
To answer that, let’s go back to the opponent-adjusted EPA power rankings. If you’re unfamiliar with EPA, it stands for Expected Points Added, and the numbers found in here all come from nflfastR’s Expected Points model.
Expected Points, the foundation of many analytical arguments, uses data from previous NFL seasons to determine how many points a team is likely to come away with on a given play based on down, distance, time remaining, and field position. The difference in expected points at the start of a play and expected points at the end is referred to as expected points added, or EPA.
In these opponent-adjusted EPA power rankings, we’re looking at the cumulative EPA for each team through four weeks, adjusted for strength of opponent. One important note: Adjusting for opponent doesn’t really increase predictiveness, but it can help add to the descriptiveness. Let’s get to it.
Last year, the story of these rankings was how good Dallas looked despite their mediocre record. They actually had the sixth-best adjusted EPA differential heading into Week 17 with a 7-8 record!
This year doesn’t look to be as kind.
They are 13th in adjusted offensive EPA, lower than might be expected, and 21st in defensive EPA, which is somehow still better than expected for the team with the most points allowed so far. They get a boost with the opponent adjustments for playing the Seahawks, Rams, and Browns, who rank 4, 5, and 6 respectively in adjusted offensive EPA.
The teams that dropped the most with the opponent adjustments added compared to their standard EPA differential are the 49ers and the Steelers, both of whom get dinged for soft starting schedules. On the other end, the Chargers and the Vikings get the biggest boost to their ranks for playing tough opening schedules.
Adding in opponent adjustments of course doesn’t always add the right context too early in the season. Football Outsiders doesn’t add in opponent adjustments until after Week 4 in their DVOA rankings.
When is the best time to add those adjustments?
We want to find a point in the season where each team has a large enough sample size to be somewhat predictive of the rest of the season. Where is that point? Is Week 4 late enough? Let’s find out.
Using every season since 2009, I looked at each team’s EPA/play on offense and defense after any number of weeks and compared it to their EPA/play over the remainder of the season. As it turns out, four weeks is about as good of a time as any to predict future performance for any given team.
A quick explainer on reading this chart: A team’s offensive EPA/play after Week 4 explains about 34% of the variance in that same team’s offensive EPA/play for the rest of the year. The correlation peaks after Week 6 and then slowly starts going down again as the number of games being predicted gets smaller, adding more variance.
So after four weeks, we’ve got about as good of an idea projecting these offenses forward as we’ll get. Defense is another story, as it’s always pretty volatile no matter how far into the season you are.
This is part of the reason you might hear the phrase “defense doesn’t matter.” Let me be clear here: defense does matter. It’s just much harder to project going forward and is impacted significantly by the offenses it has to defend against, while offenses are typically much more stable and easier to project. So what “defense doesn’t matter” really means is that we can’t make defensive projections for the rest of the year with the confidence that we make offensive projections.
Looking forward, we probably have a decent idea of how good the Cowboys are. Fortunately for them, seven of their final twelve games are against teams in the bottom six of these rankings. Dallas has a good lead within the division in terms of their EPA differential, which aligns with the fact that they’ve still got the best odds to win the NFC East at 53%, per FiveThirtyEight.
Dallas will get a golden opportunity to take over the division lead to align with those chances on Sunday when they go up against the struggling New York Giants.
6 amazing, impressive or depressing Cowboys-Browns, quarter-pole stats
[vertical-gallery id=655702][lawrence-newsletter]