Dallas Cowboys: The importance of Expected Points Added (EPA) in 2020

Ezekiel Elliott #21 of the Dallas Cowboys (Photo by Ronald Martinez/Getty Images)
Ezekiel Elliott #21 of the Dallas Cowboys (Photo by Ronald Martinez/Getty Images) /

All Dallas Cowboys fans should know and understand the value of EPA because there’s a good chance it drives the offense in 2020.

With the new Dallas Cowboys coaching staff finally came the acceptance of analytics. Rebuked by Jason Garrett and company for years, Mike McCarthy has pledged his allegiance (OK that may be stretching it) to facts and odds in 2020.

In Mike McCarthy’s well publicized “year away”, he opened himself up to new ideas and dove into the growing trend that is analytics. He even visited Pro Football Focus on his reeducation tour. As anyone who knows this genre of stats can attest – it’s a life changing event. It changes the way you personally value plays because it uses historical outcomes and odds to tangibly apply numerical value. There’s suddenly facts to replace feelings.

Leading the way in this analytics trend is Expected Points. While EP is nothing new (and if you’re a regular reader of mine you’ve heard me singing its praises for roughly four years) it’s finally gaining mainstream acceptance as of late. Teams are building and expanding analytics staffs, coaches are using them for in-game decision making, and old football guys who refuse to adapt are getting exposed.

So let’s take a moment to clarify a few things regarding EPA so we can refer back here throughout the season. Brian Burke, the godfather of football analytics, details it all here at Advanced Football Analytics. Anything he produces is must-read material and this is the ground work in which much has been made.

Why do we have EPA?

Let’s say you ran the ball for a 7 yard gain. Good, right? But what if I told you that run came on 3rd and 9 and it resulted in a punt? Then it would be a pretty terrible play.

Now let’s say in a different situation you ran the ball for a 2 yard gain. Many people would call that a disappointing run. But what if I told you it was on 4th and 1 and that run resulted in a first down and pushed the team into field goal territory. Now that’s a pretty fruitful play, right?

That’s why we have EPA. It’s applies context. EPA looks at historical data and details what amount of points can be expected from that position on the field given the down, distance, and time on the clock.

The EPA (expected points added) tells us how a play increased or decreased in expected points. So that play above that converted for first down and moved the team into scoring territory is going to be worth more in EPA than that 7-yard run and even more than a TD run on 1st and goal from the 1-yard line. It added the most to the offense’s scoring odds.

How do we use EPA?

We use EPA to judge the worth of a given play. Often times a 4-yard gain on first down actually decreases the offenses expected points (that’s a negative EPA). Since the idea is score points on offense, every play should be an effort to increase the chances of scoring.

More from Dallas Cowboys

By repeatedly running the ball for 3 and 4 yard gains on first down, we’re repeatedly hurting the offense’s ability to score. We’ll get into specifics of the Dallas Cowboys offense the next few weeks, but the goal now is to lay the groundwork for understanding that not all positive gains are good and plays that have poor success rates should be avoided. Last offseason we ran a similar exercise and broke down the results here: What EPA tells us about the Dallas Cowboys  offense.

This season we’ll do the same looking at downs and game situations. From there we can determine what play-calling is good and what is not.

Bad Stats

Passer ratings, yardage totals and competition percentages can all be misleading because they don’t have context. You can win in all of those categories without ever throwing a single pass beyond the sticks. That’s a problem.

Rushing totals, yards per carry, and positive gains say nothing about the value of the runs. You can gain 100 yards at 4ypc with zero runs for loss and still have a negative EPA for the day. It actually happens all the time in the running game. That’s why we always try to qualify these “bad stats” because on their own they say very little about the success of the player.

Assessing the Rangers Draft: Timelines for draft picks. dark. Next

Expected points added is a way to objectively grade the value of a play. It can tell us which plays actually helped the offense increase their scoring chances on the drive and which plays hurt the offense’s scoring chances. Understanding this tells us what play calls are historically wise and what calls are historically foolish.

  • Published on 06/24/2020 at 11:01 AM
  • Last updated at 06/24/2020 at 08:51 AM