Professional football is a notoriously difficult sport to predict. The 16-game schedule means fits of good or bad luck that would dissipate across an 82-game or 162-game schedule in other sports might be a team's entire season. The obvious importance of the quarterback, the most valuable position in sports, can allow a team to sink or swim on the availability and performance of one player. Roster size and attrition rates make it more difficult to rely upon the previous year's performance as a measure of what will happen next. Football will always be more resistant to quantitative analysis than the other major American professional sports.
Even with all that being true, there are some valuable underlying measures of performance that exhibit predictive value for the future. Many of them are concepts that popped up in other sports before being successfully applied to football. None of them is foolproof, but these measures often heavily suggest that a team will either improve or decline the following year.
As training camps begin to kick off and we start getting back in the swing of football, let's run through these numbers for 2015, apply a bit of real-world common sense to the factors they can't account for and see what they tell us about what will happen in 2016.
As is true in other sports, we know that a team's point differential is a better indicator of future win-loss record than its actual win-loss record. We can produce an "expected" win total for each NFL team, given its point differential, by running the Pythagorean Expectation formula -- Points For2.37 / (Points For2.37 + Points Against2.37) -- and multiplying it by 16, for the number of games in a season. Ever since Bill James created this for baseball and Houston Rockets general manager Daryl Morey modified it for football, the results have shown that winning more games than your Pythagorean Expectation tends to mean a team will decline the following season, while falling short of expectations tends to mean a team will improve.
Interestingly enough, last year the teams that were most likely to decline after outperforming their numbers in 2014 -- Arizona and Cincinnati -- both improved, thanks to much better quarterback play. The teams just behind them in the top five were Detroit, Dallas and Pittsburgh, each of which did decline. On the flip side, there were six teams that underperformed their Pythagorean Expectation by one win or more last season, and of those six, only the Giants (who maintained a 6-10 record) failed to improve. The other five teams (Tampa, Atlanta, Kansas City, St. Louis and Tennessee) improved by an average of two wins each.
The 2015 Pythagorean Expectation numbers suggest that the two Super Bowl teams exceeded their expected wins the most. That shouldn't necessarily be a surprise for the Panthers: Teams that win as many games as Carolina did last season are always outperforming their Pythagorean Expectation because it's virtually impossible to have the point differential of a "true" 15-win team. (The best point differential since the AFL-NFL merger belongs to the 2007 Patriots, who went 16-0 with the Pythagorean Expectation of a 13.8-win team.)
Too many things have to go right to make that happen again, even with a talented roster. Of the 23 other teams that won 14 or more games since the league went to a 16-game slate in 1978, 21 declined and zero improved. The teams won an average of 10.7 games the following year. Only the 1990 49ers and 2003 Patriots managed to maintain their previous records. The Panthers will be good in 2016, but it's reasonable to expect them to decline. The Broncos, who underwent a lot of roster turnover and have major questions at quarterback, will find it tough to win 12 games again. Let's just say a prayer for the 49ers, who are being described as lucky for the first time in several years.
The team most likely to improve in 2016 is the Chargers, who fell short of their expected win total by almost exactly two full wins. Since 1989, 62 teams have fallen in the range of underperforming by 1.5 to 2.5 wins. Forty-nine of those teams (79 percent) won more games the following year, and the group as a whole improved by an average of 2.3 wins. The Seahawks' fitting into that bucket might also be scary for Cardinals fans. Although Arizona won the division by three games, the Cardinals' Pythagorean Expectation (11.9 wins) was just narrowly ahead of Seattle's (11.7 wins).
Record in close games
Closely correlated with outplaying or underperforming point differential in football, the vast majority of evidence suggests that teams are unable to win a high percentage of games decided by seven points or fewer year after year. There have been a few exceptions -- teams led by Tom Brady, Peyton Manning and Andrew Luck have grossly outstripped expectations -- but generally, teams that win 75 percent or more of their close games in a given season struggle to match that performance the following season.
If you think great quarterbacks are immune to this problem, consider the plight of Aaron Rodgers. In 2014, his Packers went 5-0 in one-score games. Last year? They went 2-3, and even that required a Hail Mary to beat the Lions. Those same Lions were a classic measure of how performance in close games can swing. In 2014, they went 6-1 in one-score games and made the postseason. They went 4-4 last season and finished 7-9. In all, eight teams won 75 percent of their one-score games in 2014, finishing a combined 35-7-1 in such contests. Those teams regressed to 34-25 in close games last season.
Meanwhile, the Bears, Giants, Buccaneers and Titans -- all of whom won 25 percent or fewer of their close games in 2014 -- went from a combined 2-19 to 15-23 in such contests last season. The Bucs paced the way, as they improved from 1-8 to 5-3 in close games. The Giants and Titans didn't really hold up their end of the bargain.
Notice the overlap between the first two tables? The Panthers have seen both ends of this. Carolina started 2-14 in one-score games under Ron Rivera, but close wins started appearing midway through the 2013 season, as Rivera started to get aggressive in tight situations. The Panthers won their five subsequent close games that season to finish 5-2 in one-score contests. That suggested that they would decline in 2014. Sure enough, Carolina fell to 4-2-1 in such contests, and even that required a missed chip shot from the Bengals for a tie in overtime. It's tough to see the Panthers (or a Mark Sanchez-led Broncos team) being that good in close games again in 2015.
Although the Browns look likely to improve based on their ugly performance in one-score games, I would be skeptical that they'll gain as much as other teams that were similarly bad in this metric, if only because they've continued to set their roster back while rebuilding this offseason. In other words, Cleveland won't be as bad in close games, but the Browns might not play many close contests.
These numbers -- and more I'll mention in a moment as an even stronger indicator -- suggest the Cowboys are a classic turnaround team. The Giants, meanwhile, are hoping better performance in close games and improved health will get them back in the playoff picture. (They have been the most injured team in football three seasons running.) The Chargers went 9-7 in close games during Mike McCoy's first two seasons at the helm, so there is little reason to think they'll be as bad in those games in 2016, especially after they invested heavily in what should be an improved defense.
Creating takeaways and avoiding giveaways are skills, but players and teams can't always turn those skills into top-tier production on an annual basis. Tom Brady is one of the best quarterbacks in league history when it comes to avoiding mistakes, but his play from 2009-11 indicates the volatile nature of interception totals. He followed his 13-pick campaign in 2009 with a historic four-interception season in 2010 before reverting to 12 picks in 2011. He was the same quarterback with the same skill set those three seasons, but the players surrounding him changed, the context shifted and he was unquestionably subject to some level of luck.
With that in mind, although a team featuring Brady at quarterback is always more likely to deliver a superior turnover margin than one with, say, Eli Manning under center, teams with dramatically high or low turnover margins often regress toward the mean the following seasons, which, naturally, impacts their win-loss record in the process.
Indeed, the Packers (plus-14), Texans (plus-12) and Patriots (plus-12) were the three teams in the NFL that posted double-digit positive turnover margins in 2014 ... and their margins declined by an average of seven turnovers last season. The story was even starker for the five teams that had turnover margins of minus-10 or worse last season. Although the Titans actually declined by going from minus-10 in 2014 to minus-15 last season, the other four teams in the bunch -- the Saints, Jets, Raiders and Washington -- posted positive turnover differentials and improved their totals by an average of 16.2 spots in the rankings. As a unit, this five-team group improved by an average of 3.2 wins. Not bad.
The two outliers at the top of the charts loom large as teams likely to improve and decline. The Panthers, with much of the same personnel in key places, posted a turnover margin of just plus-3 in 2014. Their key addition in terms of creating takeaways was safety Kurt Coleman, who picked off seven passes last season after intercepting 10 in his first 74 NFL games. Coleman probably won't pick off passes at quite that rate this season. Since 1989, the 50 teams with turnover ratios between plus-15 and plus-25 posted an average margin of plus-2.7 the following season, declining by an average of 15.8 turnovers.
That's even truer for the Cowboys, who underwent a fascinating change between 2014 and 2015. Although you would figure that Dallas' disastrous turnover margin last season could be chalked up entirely to Matt Cassel & Co. at quarterback, that's only part of the story. The Cowboys' offense posted the league's worst turnover rate on a per-possession basis last season, but the team was 23rd in that category with Tony Romo around in 2014.
Dallas' defense, meanwhile, fell much further. In 2014, Rod Marinelli's unit led the league by forcing turnovers on 17.2 percent of opposing possessions. Last year, with arguably better personnel on defense, besides missing injured CB Orlando Scandrick, the Cowboys posted a takeaway rate of 6.0 percent, the worst rate in football. They should be somewhere in the middle in 2016, which will make life easier for the defense. Since 1989, teams that posted turnover margins between minus-25 and minus-15, as the Cowboys did a year ago, produced an average turnover margin of -0.1 the following season, improving by 18.6 turnovers. They also won 3.1 more games the next season. Better luck in close games and an improved turnover rate on defense should represent a notable improvement for the Cowboys, and that's before factoring in the likelihood of healthier seasons from Romo and Dez Bryant.
Defensive TDs allowed
Although creating and avoiding turnovers are skills, returning those takeaways for immediate touchdowns is not. There is too much randomness and there are too many variables involved in what turns into a scramble play just about every time. As a result, points "allowed" off defensive touchdowns often regress toward the mean the following season.
The Rams, for example, allowed a staggering 48 points on interception and fumble returns in 2014, but even with Nick Foles and Case Keenum at quarterback, that figure fell to 18 points in 2015, which was around league-average (17.4). Three of the other five worst teams in this metric in 2014 -- the Falcons, Saints and Raiders -- also allowed 18 points or fewer on defensive return TDs last season. The Ravens, Dolphins and Steelers didn't give up a single touchdown return on a takeaway in 2014, and while the Steelers managed to repeat that feat in 2015, the Ravens' total rose to 30 points. (The Dolphins produced a more respectable 12-point figure.)
The Jaguars ranked among the NFL's worst teams at allowing points off defensive touchdowns for the second consecutive season, as they upped their total from 30 in 2014 to 42 last season. You can understand why that would happen in a pass-happy offense led by a young, aggressive quarterback in Blake Bortles, but the Jags had two botched snaps and a failed toss to backup halfback Corey Grant that resulted in touchdowns. That stuff isn't likely to recur. As for the Steelers, their streak of games without allowing a defensive touchdown stretches back 45 games (September 2013), when the Bears struck twice against Pittsburgh's offense.
Losing with a halftime lead
Good teams hold their leads, of course, but even struggling teams are doing something right if they get out to hot starts and lead as games go into halftime. It seems like the sort of problem that would recur season after season, but that hasn't happened. Since 1989, teams that have blown (the admittedly arbitrary measure of) four or more halftime leads in a given season have blown just 1.7 such leads the following year. Those 81 teams have improved by an average of 2.2 wins the following season.
Last year this wasn't very useful. The Bucs, Titans, Rams and 49ers got better at protecting leads -- they combined for just eight second-half collapses after 19 the previous year -- but improved by an average of only 0.8 wins, as the 49ers dragged the rest of the group down. Similarly, the Broncos, Packers, Texans, Raiders and Seahawks, who didn't blow a single halftime advantage in 2014, saw their combined total rise to six in 2015, while their collective win-loss record stayed the same.
In 2015, the list of teams that failed to lose a game after leading at halftime expanded to seven: The aforementioned Packers and Texans stuck around and were joined by the Panthers, Dolphins, Vikings, Jets and 49ers. We'll see if they continue to pull that off. On the other side of the spectrum, the Cowboys blew five halftime leads, including an 11-point advantage over the Falcons in Week 3. The Giants and Chargers failed to hold onto four each. These three teams, along with the Titans, keep popping up as strong candidates to turn things around in 2016.
Strength of schedule
A team's schedule isn't entirely random -- the 49ers are stuck in a division with the Seahawks and Cardinals for the foreseeable future -- but as many as 10 of a team's 16 games each season can be new opponents. That can allow for some dramatic swings in schedule quality. Although strength of schedule based on last season's win-loss record is basically useless, I generate a better strength of schedule for the prior season by looking at each team's set of opponents and calculating their average Pythagorean Expectation in games not involving that team. For the 49ers, the strength of schedule calculations look at how each opponent performed in each of its games over the course of the season, excluding games involving the 49ers.
We can compare each team's schedule rank by that measure to the estimated 2016 strength of schedule generated by Warren Sharp, which uses the over-under measures set by Vegas before the season to estimate team quality. The 49ers, who had the toughest schedule by my measure in 2015, again have the toughest schedule per Sharp and Vegas in 2016.
Here are the biggest gaps between 2015 strength of schedule and 2016 estimated strength of schedule:
Although Atlanta had to face the 15-1 Panthers twice last year, their other 14 games came against opposition with an average Pythagorean win expectancy in non-Falcons games of 0.410; in other words, the Falcons' typical non-Panthers game was against a team that would typically go 6-10 or 7-9. They managed to win only seven of those 14 contests. Last season, Atlanta faced the NFC East and AFC South, plus placement games against the Vikings and 49ers; this season, they'll face the AFC West, the NFC West and, by virtue of finishing second in the NFC, the Eagles and Packers. Not as fun. As Chase Stuart noted, the Jets are projected to be underdogs in their first six games. As far as schedules getting easier ...
Shareholders who made it to Lambeau Field recently will be happy to hear that the Packers will face a significantly more comfortable schedule this season. While I am sure Aaron Rodgers & Co. would have preferred to win the North last year, their second-place finish makes for an easier slate this time. Instead of the Cardinals and Panthers, the Packers will suit up for placement games against the Seahawks and Falcons. They also swap eight games against the NFC West and AFC West for the friendlier confines of the NFC East and AFC South.
Again popping up as a team with better luck in 2016, we see the Cowboys possessing what Vegas believes to be the easiest schedule in football.
In all ...
Last year, the numbers seemed to suggest that the Packers and Lions would decline, which I theorized might open up the NFC North for the Vikings. The Bucs were a classic improvement candidate, and though they didn't suddenly climb atop the NFC South, they tripled their win total from the previous season. The numbers also seemed to point toward the Cardinals and Bengals declining. As always, these are useful indicators that will point us in the right direction a reasonable amount of the time, but a career season from a quarterback can throw everything off.
This year, the underlying metrics point toward the Panthers and Broncos declining. It's hard to make it back to the Super Bowl, and the numbers reinforce that. Meanwhile, the Cowboys and Chargers have relative mountains of hard evidence suggesting that they played better than their records in 2015 and are likely to take a step toward contention in 2016. It might be too much to ask for those teams to go worst to first, but along with perhaps the Giants and Titans, they're the teams we should be looking out for as fast risers this season.