Upset Watch: Super Bowl LII
2 Feb, 2018 by Shaun Lowrie
This week's article is a long one, in which we'll cover some plans for the site, the statistical basis we're using this week, and of course, tell you who we think will win. That's it. I even got through the whole season without mentioning that I got a kitten in Octo... oh. Sorry.
Changes Coming in 2018:
This offseason will see the site undergo the biggest changes in our 6 year history that will give a hugely improved experience on Pickwatch.
I always envisioned Pickwatch as a wide-ranging resource for people to use picking games, a site that caters to every ability and allows those without any statistical background to understand the context of the information you are viewing. I know the expert data is why people come to Pickwatch at the outset, and I can assure you we'll always keep that at our heart, but we've always strove to add more depth to the statistics we use, because we know that like us, you're football junkies who love the analytics that help you win. At some point, that same ethos began to morph into something tangible as a new feature for Pickwatch that will offer levels of depth and automation not currently available to those who predict games.
Our aim for next season is not just to write about our advanced data in small doses, but for you to be able to harness it yourself for any game in any week, on demand. Our new platform will not only mean an overhaul of the way you interact with the site in order to make this something you can access on your phone (and by app...) but will mean that nobody need ever export data to an excel document again. You'll have a large array of statistical tools and filters at your disposal that don't just refer to expert data, but also real life trends and statistics from the NFL. You'll also be able to make your own ATS picks for the first time, and track your own performance in far more depth than you can now.
So I just wanted to let you know before we end this season, that for us, the next 7-8 months are going to be the most intense work we've done yet. We're overhauling every element of the site to be faster, smarter, and more in-depth.
Now that's out of the way, there's something else going on this week, and I can't remember what...
Upset Watch: Super Bowl LII
Super Bowl History
The first question we're going to ask is the most pertinent: How often do teams upset the favorites in the Super Bowl? The answer? Very often indeed. In fact, so often does it happen, that favorites have not won back to back Super Bowls in over a decade, when the Steelers beat the Seahawks in 2006, and the Colts trounced the Bears in 2007 (the prior two Super Bowls had also been won by favorites in the Patriots).
In the following 10 Super Bowls, underdogs have won 6 of the 10 games. That includes two occasions when the Patriots were upset by the Giants under Tom Coughlin and Eli Manning.
This isn't something that should immediately signify an Eagles win, but it's worth considering that upsets on this stage are actually incredibly common, and of course, in their nature, rarely predicted by a consensus. These are the two best teams in the NFL, and to get to the Super Bowl, it's unusual for them not to have beaten at least one marquee team in their conference, so both teams should realistically be objectively there on merit. More on that later...
What about the ATS record in that time?
Well boy am I glad you asked.
With so many underdog victories, we know that favorites aren't covering very often. but what is remarkable is how often they do cover when they win. Again, just one Super Bowl in the last decade has ended with a favorite winning but not covering (the Cardinals ran the Steelers close in 09 before the Santonio Holmes circus grab to win it).
That spread was -6.5 in favor of the Steelers, and there has not been a higher spread since. In fact, the opening line in this year's game (-5.5 to the Pats) would have been the highest, however that came in to 4.5 and as low as 4.
What does this tell us? Well, statistically, our best ATS bet is likely the team we think will win. The chances of it being a Pats win and an Eagles cover are very slim indeed. That's not to say it can't happen (all patterns are broken), but it's something to be aware of before assuming it will be close. The more that line comes in, the more that rings true.
Analyzing the Teams
It's tempting to call the Eagles a team of two seasons. When Carson Wentz tore his ACL in week 14, I was actually quite positive about his replacement Nick Foles, picking the Eagles to beat the spread against the Giants, which they promptly did.
They say, however, that familiarity breeds contempt. Ok, maybe contempt is a strong word, but the reason that the Eagles were such big underdogs in the Divisional Round against Atlanta was that Foles was deeply unimpressive against Oakland and Dallas in the end-of-season run-outs he got.
Foles proved us doubters wrong, however, by engineering not only an upset against the Falcons, but a staggeringly clinical shellacking of the best defense in the NFL - Minnesota - in the NFC title game. His performances in those two games have made picking the Super Bowl incredibly difficult, but in simple terms, here's the one lesson I think is important:
Judge the Eagles as a team
Before I started my statistical research in our database of predictive information, I had an idea in my mind: 'The only games that matters are the ones that Nick Foles started'. That was my mindset, that the differences between Wentz and Foles were so great that it was useless to compare the Eagles of the first 13 games with the Eagles of the last 5 games. But the more I considered it, the more I knew that was incorrect.
First off, we assume Wentz was better than Foles, which is correct, but to what degree. Foles is not a scrub, and in fact he and Wentz have almost identical passer ratings throughout their careers (Wentz 88.7, Foles 87.4), so the drop-off is not as acute as we might assume.
Secondly, and perhaps most tellingly of all, Foles has actually played incredibly well in 3/5 starts, with his worst game being the end of season dead rubber against the Cowboys, which perhaps skews his stats unfairly given the eventual outcome flattered neither team as entire units. Foles has 3 games out of the other 4 starts where his passer rating was in triple digits, and his TD to INT ratio for the season is 8:2.
But again, this is really just the foundation, because what our analysis of Foles is really telling us - with the greatest of respect to Foles - is that we can ignore him. Yes, Foles is good enough that we do not have to compartmentalize the Eagles' season into two sections, which is good news for all of us, because our sample data is massively increased.
The Eagles led the league in run defense, averaging just 79.2 yards allowed to opposing rushers. Given how much the Pats rely on the pass, it'd be tempting to dismiss or at least discount how much this will effect the game, but again, you'd be dramatically wrong. New England lost 3 times in the regular season, and in each of those games, their leading rusher did not break 50 yards for the game. The Patriots are good, but they are still a football team, and football teams run the ball when they can. Against the Eagles, running the ball is as hard as it comes.
Here's a quick breakdown of something I want to show you about the Eagles that gave me a huge pause for thought:
These games were against the best defenses that the Eagles faced in 2017, with the Patriots in there for comparison. As you can see, the best defense they played was the Vikings in the NFC title game, giving up just 15.8 points per game, and the Eagles outperformed that average by 22.2 points. Only once did they fail to outscore their opponent's average points allowed.
Perhaps most interestingly, they have also been expected to lose 4 of the 6 games. Guess what?
They didn't lose a single one.
Yes, this Eagles defense is elite, but the offense? You'd better believe it's a huge reason that this team wins when they are underdogs. The Pats are one of the best defenses in the league, of that there can be no doubt, but they're actually the exact kind of defense that the Eagles make a habit of upsetting.
New England Patriots
The Patriots are a Super Bowl history of their own. We all know the games of the Brady-Belichick era by now. The Rams upset, John Kasay's out of bounds kick-off, THE COMEBACK... but there are also two Super Bowl losses in there to the Giants in the last 18 years (good god, it could have been 8...) and they are not infallible just because of their experience on this stage.
But before we talk about things that might go wrong, let's talk about what's gone right. The Patriots are almost a twin to the Eagles (hey, there's a Minnesota-St. Paul reference there right?) in that they scored just 1 point more in the entire regular season than Philadelphia. Both teams, therefore, average 28.6 points per game.
This will be important.
The Pats are extraordinarily good at outmatching defenses. failing to outscore their opponents average ppg allowed just twice all year. This is an offense that is based on exploiting any and every weakness a defense can give them. They have a pass-happy attack (the biggest difference between the two teams) and as mentioned, don't rely on getting Dion Lewis or Rex Burkhead going to win. That could well be important against the dominant Eagles defensive line.
As for their defense? It's massively improved since the early weeks of the season, when people jokingly suggested that the Patriots D would cure any offensive woes their upcoming opponents may have. Well, those people may now be laughing on the other side of their face (weird), because here's what they've done:
Since week their week 4 loss to Carolina, the Pats have only failed to hold opponents under their average points per game once. For comparison, the Eagles allowed 9 opponents to score more than their average points per game this season, so that stretch of games is quite remarkable. But how much does it actually mean in the context of this game?
My initial thought was this: Statistically, I'm looking for the winner, and the Patriots? Well, they're definitely winners, right? My inclination was to pick the Pats to win another championship and for us all to wonder where this ends. Maybe in awkward questions about where Brady might keep the 11th ring he wins. I don't know, I'm spitballing here...
But I have some problems here. The stats are not really in their favor at all. Sure, they're a great team, but are they better than the Eagles? More important to me is not that both teams win (of course they do), but how do they win?
Super Bowl LII Pick: Philadelphia Eagles (+4.5)
We're picking the dog, and here's why...
I don't think the Pats are a match for Philly in the departments that matter most. Let's take the offenses. The Pats, as we noted above, roundly outscore their opponents and routinely perform above expectation. The problem? They still lose when they do this. More importantly, again, the Eagles have allowed teams to score more than their average allowed nine times in 2017 - yet they have lost only two of those games. Conversely, the Pats lost 3 games in 2017 and outscored their opponents' average in two of them.
In short, the Eagles give up points and still win, while the Pats can score big points and still lose.
Perhaps also in play is the fact that the Pats are one of those aforementioned 'unusual' teams who have quite frankly, had a pretty easy route to the Super Bowl, beating Tennessee in the Divisional round and Jacksonville in the AFC title game. Not to denigrate the Jags, who I believe are a great defense, but the Patriots struggled mightily to overcome a team that is deeply flawed at QB and whose strengths in the run game are exactly in keeping with the strengths they're going to face on Sunday against the Eagles.
The Eagles offense has a propensity to outscore the best defenses very heavily even when they are underdogs. I always talk about big game mentalities at the Super Bowl, because how teams react in a given circumstance is a good sign of how a team is built. Underdogs who find themselves favorites in the Super Bowl are often thrown into a strange circumstance where their 'nobody believes in us!' mantra is simply not true, objectively. Well, for the Eagles, that mantra will certainly still be true. They find themselves in a situation uncannily like the previous two playoff games they've won, and perhaps most interestingly, they face an opponent whose strengths meet all of the criteria that the Pats will dread.
I said earlier that the team you choose should be the one you take vs the spread, and I stand by that. If you believe in the Pats, you should take them at the current -4.5, as it is not high enough to rethink the last decade of statistical precursors that suggest it is unlikely they will win without covering.
But for me, it's the underdogs. It's more fun that way in any case...
Shaun Lowrie is the founder and editor of Pickwatch, a site designed to give you a new perspective on experts in the NFL media by tracking their performance when picking games or predicting draft picks, team performance, or player potential. You can contact him at [email protected] or on twitter at @pickwatch