Which stats matter most for the Super Bowl?

We inch ever forward to the Super Bowl and the smashing of pads between the Patriots and Eagles. You flip through the channels and the hum of football analysis grows until that deafening cheer as the ball is kicked off to begin the game where one team stands with the Vince Lombardi trophy lifted high. Who will win? What should we look at? What stats really matter?

Let’s turn to an ESPN article appropriately entitled The NFL stats that matter most by Bill Barnwell. Let’s see how these statistics stand up over several years of NFL data and how our championship teams fared against these metrics.

Let’s look at the stats.

DVOA (Defense-adjusted Value Over Average)

This stat is given as percentage and measures a team’s success on a given play versus the expected performance over time. A team with a 10 percent DVOA is that much better than the league average on a play-by-play basis.

How do past Super Bowl teams perform? At the bottom of this article, you can explore our interactive tool made with Tableau. For now, here are our insights. First, three of the past five Super Bowls featured the team with the highest DVOA in the regular season. The 2013 Seattle Seahawks and 2016 New England Patriots finished the regular season with the highest DVOA, and went on to win the Super Bowl. However, this was not the case for the Seahawks in 2014.

How about this year? You’ll see the matchup between the Philadelphia Eagles and the New England Patriots as the narrowest margin in DVOA between Super Bowl opponents in the last five seasons.

Point differential

The Pythagorean expectation is a measure of how many games a team should have one and even how lucky a team appears to be if they’ve won more games than they should have, under this metric. Some say it isn’t luck — it is coaching! Could it be that the Patriots or Eagles are lucky enough to outperform their stats? Are one of the teams about to have their luck run out?

First, let’s peer back. Notice AFC Super Bowl teams in the past five seasons have finished the regular season “luckier”, meaning they won more games than Pythagorean expectation would have predicted. The only time this wasn’t the case was in Super Bowl 50, a matchup of the two luckiest teams in football that season.

So, what about those Eagles and Patriots? Both rank in the top 10 luckiest teams this regular season, but both teams won only one game above Pythagorean expectation. One team’s luck will run out very soon.

More gridiron stats

Let’s turn to quarterback statistics. There are several that you’ll find. First and most straightforward is yards per attempt. There is also Total QBR which, while having flaws, gives a one-number metric to compare quarterbacks. So, how about Brady and Foles? How much data do we have on Foles and what does it say?

In the Tableau application, you’ll see a comparison of Brady and Wentz. Isn’t that the conundrum of the playoffs? How will the Eagles do with Foles continuing to take snaps? Time will tell. You have much to explore with statistics like Catch Rate, Running Back Success Rate and QB Hits. Are these the stats that really count? Here is your dashboard to explore the stats and come up with your own analysis! Which stats do you think matter? Which might give you more insight on what we’ll be talking about Monday morning?

In the end, explore the stats, look for patterns, be your own sports analyst and find an insight that you can share at a Super Bowl party that might momentarily silence your friends amongst the inevitable clamor of the big game.


This is article is co-authored by Scott Teal, Sr. Data Artist at Tableau, and Dr. Tim Chartier. This is part of the Cats Stats series on using ESPN in the Curriculum.

Leave a Reply

Your email address will not be published. Required fields are marked *