The expected wins model is one of the most common measures used by sabermetricians to calculate teams win percentage over the course of the season. The stat and formula was developed by the godfather of advanced stats, Bill James, who first used it in baseball.
With less than a few days until the beginning of the 2016 NFL season, we decided to dig in and find out for ourselves what the projections say. Some of the notable regression and progression candidates might surprise you.
This formula has had an incredible amount of success, especially with outliers who have over a two-win differential in their Pythagorean win expectations and their actual record from the previous season. Candidates to improve, with a differential of over 2 wins, have averaged 2.7 more actual wins the following year. While regression candidates with at least two actual wins greater than their projection, have come back to earth winning an average of 2.1 fewer games. That's over a 35-year span.
In this year's numbers, it’s the two conference champions who have a differential greater than two wins. While the one improvement candidate who has a two win difference in the projections is San Diego with 1.97 change.
As our first piece of many, digging deep into the stats to figure out ways to beat the spread, here’s what we found.
The following are contextualizations of the data that you can find in the chart below.
For starters, it’s not easy to repeat a 15 win season in the NFL. The six teams in NFL history to go 15-1, prior to Carolina, have regressed at an average of 3.84 wins the following season. Carolina is also dealing with a Super Bowl hangover and the loss of Josh Norman.
Even more importantly, repeating an almost perfect record his not an easy task. The 1986 Chicago Bears are the closest to repeat the feat by finishing 14-and-2 in the regular season.
The Colts are a tricky one because the numbers don’t take into account the seven-game absence of Andrew Luck in 2015. Though it’s important to remember that the Colts franchise quarterback only went 2-5 in those seven starts.
Luck might be part of the Colts problem, who defied all odds by winning 8 games in 2015, as his contract extension has severely lessened Indy’s depth. It's hard to find many improvements on this roster that lost some significant defensive starters, such as Jerrell Freeman. Even with a healthy Luck, this team might not bounce back into playoff contention like many expect, in a more competitive AFC South.
Things in Chargers country did not go well last year, as they won a mere 4 games and faced several injuries throughout.
Ken Whisenhunt is back coaching up Phillip Rivers, who played some of his best football, last time under his former offensive coordinator. The Chargers are also a young team who seem poised for a bounce back season, with their total wins projecting out to 6 in 2016.
A young defensive core, with several players who’ve shown promise, should help significantly in San Diego exceeding their miserable season.
Another young team that faced injury issues and has made significant improvements in the offseason. Again, this isn't accounted for in the numbers but helps contextualized them. A major factor for Tennesse was close losses, something that should improve with a more experienced young core.
The Titans were 2-and-6 in one possession games, with four of those losses coming by three points or less. With a healthy Marcus Mariota in year two, a promising stable of running backs, added depth, and talent across the roster, the Titans should be able to exceed their 3-win total from 2015.
Seattle’s projection comes as somewhat of a surprise, as they're not your typical improvement candidates, like two teams above. However, the Seahawks “only” won 10 games in 2015. They also had a very rough beginning to the season in which they went 2-4, while dealing with issues such as the Kam Chancellor holdout and a bit of a Super Bowl 49 loss hangover. Seattle’s rough first month and a half of the year was also a product of two overtime losses.
Seattle’s loss of Marshawn Lynch shouldn’t be a huge factor, as the running back only played 28-percent of snaps in Seattle. Russell Okung and even Bruce Irving are slightly more significant departures, having started over 70-percent of snaps. A veteran roster that’s used to winning should overcome the few losses and seem like a healthy bet to exceed the 10-wins from last season.
The Giants had a very high number of one-possession games, 11, and a 3-8 record to go with it. New York changed their head coach but maintained the rest of their staff, ensuring some sort of stability in the 2015 figures translating to this year.
They’ve also made some splashy free-agency additions and seem primed to make even greater strides in year two in Ben McAdoo’s offense. On the defensive side, the Giants have spent the aforementioned money and should at least improve from their third worst points allowed in the NFL from 2015.
Below is the entire chart with the expected win numbers and differentials from 2015 for every team.