NFL Stat-Around: Who Will Win the Super Bowl?

Heading into the playoffs, multiple teams can make a claim that they are the favorites to win it all, and even those who just barely snuck in have hopes. We decided to look at several of the playoff teams from an analytical perspective, drawing attention to a stat or two that supports the idea of them hoisting the Lombardi Trophy come February 7th.

Super Bowl Champions: Seattle

By Kurt Bullard

 This year, the Kansas City Royals proved that the curse of the runner-up is not real, bringing home the World Series only one year removed from a heartbreaking Game 7 loss to the Giants in last year’s Fall Classic. I wouldn’t be surprised if Seattle followed suit.

The biggest issue facing the Seahawks in the first half of the year was its lack of protection of Russell Wilson on the field. The defense is still elite—the team finished first in terms of points allowed on the year—but Wilson found himself on his back a total of 31 times in the first seven contests. Consequently, the team only broke 30 once, and it was in a loss in its season opener. Whether you line up Tom Brady or Josh Freeman at QB, no quarterback can regularly succeed while receiving constant pressure.

But for the last nine games, the Seahawks found a way to give its star QB more time to throw, only allowing 14 sacks in its last nine games. Since then, the results speak for themselves—notching 26 TDs to only 3 INTs en route to 4,000 yards passing and 500 yards rushing by year’s end.

The impact of the offense is recognized by most reputable advanced statistics websites. Football Outsiders has Seattle as its No. 1 team per DVOA.

With a healthy Beast Mode returning which in most likelihood only improves the team’s red zone offense and ability to eat up the clock in crunch time, the Seahawks seem primed to challenge the Cardinals and Panthers for the right to travel to San Francisco.

 

The Redskins Could Take the NFC. Stop Laughing.

By Jason Light

No seriously, stop laughing.

There are a couple easy arguments one can use to pick apart this team here. The Skins 9-7 record is not exactly inspiring to begin with. Considering that six of their games were played against members of the second-worst division in football, it’s concerning. In the two games they played against current playoff teams, the Patriots and the Panthers, they lost by a combined score of 26 to 71. With a slightly above-average offense, ranked 10th in points for, and a vastly improved but still below-average defense, ranked 20th in DVOA, the Skins have the quintessential look of a one-and-done team.

However, the Redskins picked the perfect time to play their best football. They’ve won their last four games by a combined margin of 131-93 and have gotten key players back from injury. Jordan Reed has quietly rivaled the Eiferts and Gronkowskis of the world for the title of best tight end, finishing 5th in receiving yards and second in TDs at the position. And he finished the season on an absolutely absurd tear:

Screen Shot 2016-01-06 at 7.26.16 PM

From Week 14 up through when he was benched at halftime in a meaningless finale against Dallas, Reed caught an unbelievable 94% of the passes thrown his way for almost 400 yards and 5 TDs. Unsurprisingly, this production coincided with Kirk Cousin’s drastic improvement. Take a look at his numbers for each quarter of the season:

Screen Shot 2016-01-06 at 7.27.18 PM

This doesn’t prove that Cousins is all of a sudden an elite or even a good quarterback. But he is a quarterback who is playing his best football at the perfect time, something that certain Giants and Ravens teams of the past have proved can be invaluable to a surprise contender.

Washington is not the best team out of the NFC – there’s a reason that they have the worst record of the six contenders. But Carolina and Arizona have looked mortal as of late, with ugly outings versus Atlanta and Seattle tapering their stock, and Jordan Reed would likely be able to do some damage against a Seattle defense ranked 27th in DVOA against tight ends. Green Bay, their first round opponent, put up a paltry 21 points combined with two losses to end the season on a down note. Meanwhile, with a hot quarterback, elite receivers in Reed and Jackson, and an opportunistic defense eighth-ranked in takeaways, this Washington team is suddenly formidable.

You like that?

 

Patriots are Most Likely to Get There, Cardinals are Most Likely to Win

By Harrison Chase

I’ve been advocating the use of our POWER Rankings all season, so it’s only right that I use them to try to predict the playoffs. After updating them for the Week 17 games, I then ran a Monte Carlo simulation of the playoffs. I set homefield advantage to be a 7% increase in winning percentage, as it’s been estimated to be. The results of 100,000 simulations are below.

Screen Shot 2016-01-06 at 11.55.42 AM.png

The Cardinals, despite suffering an embarrassing beatdown at the hands of the Seahawks last week, not only top the POWER Rankings but are also the team most likely to win the Super Bowl. Below them the Carolina Panthers and New England Patriots are essentially tied for second most likely.

One percentage that may surprise people is that the POWER Rankings give less than a 6% chance of winning the Super Bowl to the Seahawks, far less than the implied Vegas odds (they currently are at 6/1 odds to win it all). This isn’t because POWER Rankings think they are bad in team – they are ranked a close fourth, trailing only Arizona, New England and Carolina (in that order). Rather, the fact that they would have to play three away games, against the Vikings, Panthers and then probably the Cardinals is a tough task for any team.

Meanwhile, for Patriots fans like myself, just because we aren’t the favorite to win it all doesn’t mean we should despair. The team fielded in a week will look vastly different from the one in recent weeks, and the POWER Rankings may be underestimating their chances for that reason. In addition, the Pats have the highest chance to make it to the Superbowl of any team, narrowly edging out Arizona, thanks to a relatively weak AFC – in our simulations they make it to the big game 37.4% of the time.

 

The Arizona Cardinals are as Legit as They Come

By Austin Tymins

After running 1 million simulations using our POWER Rankings, the Arizona Cardinals are our Super Bowl favorites with 21.1% chance of winning the title with the next closest competitor being the Carolina Panthers with 19.4%. This result speaks highly of the Cardinals’ ability especially when you consider these simulations account for the gauntlet that is the NFC side of the playoffs this season and the home field advantage Carolina would have in the NFC championship.

The Cardinals have gotten to this point because of a balance on both sides of the ball that is rarely seen in the NFL. The Cardinals were the league’s top Offense this year at 408.3 yards per game while averaging 30.6 points per contest. They were also ranked as the league’s 5th best Defense while only allowing 321.7 yards per game. Analytics agrees with this conclusion as well with the Cardinals Offense ranked 4th in the league and the Defense ranked 3rd by DVOA. This type of balance is rarely seen, and when it is, it often occurs on Super Bowl-caliber rosters.

General Manager Steve Keim has done nothing short of a remarkable job on the personnel side. In the last 3 years, the Cardinals have drafted John Brown in the 3rd round, Tyrann Mathieu in the 3rd, Alex Okafor in the 4th, J.J. Nelson in the 5th and offensive ROY candidate David Johnson in the 3rd. This is in addition to the veteran free agent signings including pro-bowler Mike Iupati in the offseason and team sack leader Dwight Freeney in October. Even with the season-ending injury to the Honey Badger, the Arizona Cardinals have a dangerous combination of youthful energy and veteran experience that is hard to find in football.

Arizona’s distinct advantages come at quarterback, wide receiver and in the secondary. Pro Football Focus ranks Carson Palmer as the 2nd best quarterback in all of football this year with a 99.6 ranking only lagging behind Ben Roethlisberger. The Cardinals are the league’s best vertical passing team which complements the insane speed of their wide receivers on the outside. On 3rd downs and Red Zone opportunities, Michael Floyd and Larry Fitzgerald have put up arguably the best statistical seasons of their careers.

On the opposite side of the ball, Arizona chooses to blitz often while trusting their secondary to keep opponent passing contained. This secondary was dealt a devastating blow when Tyrann Mathieu, the league’s top cornerback according to Pro Football Focus, tore his ACL in garbage time against the Eagles. Fortunately, the other side of the field has been on lockdown under Patrick Peterson who has played exceptionally this season allowing the fewest yards in coverage per snap in the NFL, making him a legitimate Defensive Player of the Year candidate.

The Cardinals balance makes it hard to choose a specific weak point on the roster. However, critics will likely point out that the edge rush and rushing game have struggled in the past and were both identified as weaknesses early on in the season. The Cardinals are 20th in the NFL in sacks which was greatly improved by the eight sack performance in week 16 against the Packers, while the rushing game has been more successful.

Entering the season, the Cardinals were not expected to run the ball effectively. They ended 2014 second to last in yards per game and last in yards per carry. However, the signing of Chris Johnson and the introduction of rookie David Johnson have completely changed that narrative. The Cardinals, ranked eighth in overall rushing, should finish in the top 10 in the NFL for the first time since 1984. While this change is largely attributable to Carson Palmer’s ability to stretch defenses and pull safeties out of the box, it is hard to overlook David Johnson’s performance. He is fourth in the NFL in rushing DVOA of players with a minimum of 100 rushing attempts, ahead of players like Adrian Peterson and fellow Offensive Player of the Year candidate Todd Gurley.

 

The Case for Denver

By Benedict Brady

 Unlike the past few years, this iteration of the Broncos is defined by their truly dominant defense. Per Football Outsiders, Denver has a historically great defense this year. They are ranked first in DVOA, and eighth all time (since 1991). Since 1991, 21 teams have broken -20% in Defensive DVOA. Of those 21, five have gone on to win the Super Bowl. Denver’s defense has posted a -25.8% rating and lifted the team to 12-4 and the first overall seed, despite horrendous play from Peyton Manning over the first half of the season and replacement level play from Brock Osweiler over the second half. This defense alone could win Denver the Super Bowl.

Despite this impressive performance, the Broncos will not be favored to win the Super Bowl without some real offensive improvement. And the best case for this happening is with the return of Peyton Manning. During the regular season, the Broncos offense was ranked an embarrassing 25th in DVOA. While Manning is clearly not going to return to the glory of his 2013 season, he has dropped off far more than most people would have expected. After posting a QBR of 73+ every year since QBR was created (2006), he finished 2015 with 45.0.

So, why is Manning the answer? Despite posting a QBR a few points lower than Brock Osweiler over the course of the season, Manning has shown his ability to reach the heights of his former glory. He has reached a QBR of 70 three times (highlighted by his 93.6 rating versus Green Bay), something that Osweiler has only done once, even though Manning has played only two more games. In fact, the standard deviation on his ten performances this season is an astounding 28.8! This is over ten points more than Osweiler at 18.2.

High variation is not necessarily a good thing in terms of quarterback play, but in Manning’s case it highlights his potential. With a two week rest after taking the second half of the season off, it seems likely that he would be trending up. The level of variance in his performance this season means that while he can have some terrible performances, his ceiling is still very high. Additionally, it is not as if Manning needs to be the best quarterback in the league. Three of the five great defenses mentioned above that won the Super Bowl, have had offenses ranked in the low 20s. Manning just needs to be a slight upgrade.

If Manning has a few above average performances, then the rest of the offense is still intact from last year. The offensive line has only fallen five spots in DVOA (probably partially due to quarterback play), and the team has brought back many of the same skill position players. With an offense on the upswing and a historically great defense, the Broncos have a real shot at winning Manning his second Super Bowl.

About the author

harvardsports

View all posts

4 Comments

  • For Harrison Chase:

    Admire your commitment to a statistical prediction model.

    None of the (public) predictive models I am aware of prove better than the best human experts to my knowledge. Perhaps you need a different test (other than the bar of 100% accuracy).

    I manage a little professional football prediction contest that my family started two years ago. The aim is to pick three correctly each week and we humans surely struggle to correctly pick winning “trifectas” each week.

    We also have found that the statistical models prove to be no better. We are tracking a competition to see if “The Machines” help our predictions. If you would like to follow along and/or correct our use of your predictive model’s odds, please see http://www.roi2partners.com/wftl/super-turing-bowl-i/.

    An excerpt about our Turing test is below.

    Regards,
    W.O.

    ****

    Apparently, any machine or optimization model can pick games as accurately as the average human (~63%) based on perceived strength or record. The real question is: which machine models a bias most like a human for its own profit? Ultimately, if the Machines hope to swindle The Odds Makers, they must do more than pick games as accurately as humans (63% of the time). The Machines must submit accurate picks with the appearance of human bias — or risk being identified and blacklisted by The Bookies.

    To this end, the Clearinghouse has organized a Turing League and will stage the first-ever Super Turing Bowl as a test to determine which professional football prediction model performs most like biased Humans while maximizing its rewards!

  • I am a Broncos fan and a quantitative researcher … and this year’s championship team illustrated the limits of statistical forecasting. Denver suffered in most metrics due to the anemic performance of the offense. However, the team had a winning record against playoff-bound teams even with a declining Manning or a back-up in his first series of starts. In other words, they went 12-4 with one hand tied behind their back. This was evidence of a remarkably talented roster with strong veteran leadership countering this major flaw. A team like this Denver team demonstrate that the space of possible championship teams is wider than metrics like the DVOA can cover.

Leave a Reply

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