An analysis of 54 Super Bowls

Super Bowl LIX / 59 has been in the books, the king is dead, my model has been saved, and it’s definitely time to do a retro across Super Bowls to see what we’ve learned!

I run a weekly NFL ratings model that uses data going back to the 1970 and produces ratings starting with the 1971 season, meaning I have metrics on every Super Bowl team that participated from Super Bowl VI onwards. I’ve cut this database several ways, and I hope it’s interesting to read through.

First off, the age old question … what’s the greatest NFL team of all time? Well, according to my model, here are the top Super Bowl winning teams since 1971:

Notably, the 2024 Eagles slot into the 14th spot here and are the highest rated champion since the 2013 Seahawks! This was a historically great team, particularly so in the salary cap era, as they are the 6th-highest rated champion since the salary cap was instituted in 1994. Big big congrats, Iggles fans!

Now if we flip this table, here are the worst Super Bowl champions:

It shouldn’t be a surprise that arguably the biggest Super Bowl shocker of all time - the 2007 Giants beating the undefeated Patriots - produced what the model would consider the worst champion ever. Three of the next five teams on the list were from the three previous Super Bowls before this year’s. All that is to say, it’s been a while since we’ve seen greatness.

Speaking of those 2007 Patriots, what if we change our focus and look at the best Super Bowl losers?

Yep, it’s them. The Patriots also show up at #5 on this list when they lost the sequel to the Giants in 2011.

Let’s flip this table and see who the worst Super Bowl losers - and thus the worst Super Bowl teams period - are!

Well, well, well, look who shows up at #5 on this list, the Chiefs who defied conventional wisdom, every mathematical model that exists, and probably the laws of nature themselves to claw their way to the Super Bowl. Their closest comp in the model is the 1992 Bills who also made three (and ultimately four) Super Bowls in a row, but of course famously lost all of them.

Okay, what if we focus on overall matchup quality now? Let’s take a look at the teams with the highest combined ratings going into the Super Bowl. The idea being the higher that combined rating is, the better the overall quality of the matchup is:

What’s interesting here is how many blowouts we see even when both teams have double-digit ratings. But then again, the 1980s and early 1990s were notorious for Super Bowls that were over at halftime, and teams from those eras tend to dominate these lists.

The best matchups of this millennium both involved the Patriots coming back from double-digit deficits in the 2nd half to win the championship against worthy opponents with birds as their mascots (2014 Seahawks and 2016 Falcons).

Okay, time to flip it, what are the lowest quality combined matchups we’ve ever seen?

Yeah, that looks about right. I was thrilled about the 2021 Bengals, flying to Cincinnati to be in the stands for their first playoff victory in over 30 years, but this was a team that caught fire at the right time, not a truly great team. And the same goes for their opponent. Notably, a lot of recent Super Bowl matchups, including the one we just saw, show up on this list.

Okay, time for another view. Let’s look at the projected point margins for these Super Bowls and how that lines up to the actual results. First, the closest predicted matchups, where my model predicted a scoring margin of less than 3 points:

Overall my model actually correctly picked the winner 67% of the time and that makes me happy considering these should be the hardest set of games to predict!

Next up, the medium margin games, where REACT projected the teams between 3 and 7 points apart:

Not a ton to say here, but the model’s performance bumped up here, as you’d expect, and it picked 72% of these correctly.

Next up, the biggest projected margins, all games where the model projected one team to win by at least 7 points:

Overall, my model got the winner correct in 83% of these matchups, with Tom Brady’s Patriots being involved in both of the exceptions.

More importantly, this past Super Bowl was actually on the list of predicted blowouts, belying the 1-point line (with the Chiefs favored!) we saw from sports books. While there was clearly a Chiefs / Mahomes / Reid mystique factor for the past few years, it looks like for a least the latest Super Bowl … the numbers didn’t lie.

And with that, I’m done. Enjoy the offseason!