Halfway through the 2020 NFL season, we have a decent idea about who the best and worst teams are — even if most of this year’s contenders have flaws that could keep them from reaching their full potential in the playoffs. But not everything is always what it seems in a half-season of football. Some of this year’s teams are winning a lot more (or less) than their underlying numbers say they should, which could give us clues about what the rest of the season may hold.
One way to judge how lucky a team has been is to compare its record to what we would predict based on its points scored and allowed — also known as the Pythagorean expectation.1 Some teams, like this year’s Ravens, have won exactly as many games as we would expect from their point differential. But when a team wins more than the formula says it “should have,” that usually means it won an unusual number of toss-up games, which may not be sustainable going forward.
[Related: Did The Saints And Bills Show The Bucs And Seahawks Who The Real Contenders Are?]
So far in 2020, we’ve seen a few notable outliers in terms of team wins and losses being out of sync with Pythagorean records. On the high side, the Pittsburgh Steelers are a perfect 8-0 despite a .710 Pythagorean mark — which would translate to only 5.7 wins instead of eight. Among all 31 teams to have started 8-0 since 1960, the 2020 Steelers have the fourth-worst Pythagorean record, ahead of only the 2006 Indianapolis Colts, 2015 Carolina Panthers2 and 1990 San Francisco 49ers:
The luckiest 8-0 teams?
Among NFL teams that started 8-0 since 1960, largest gaps between actual and expected record (based on the Pythagorean expectation*)
|Year||Team||Record||Scored||Allowed||Pythag. W||Diff. vs Exp.|
Obviously, you’d rather be one of the worst 8-0 teams than one of the best, say, 4-4 teams. On a certain level, it doesn’t matter how you get those eight wins, just that you got them. But that 2.3-win gap between Pittsburgh’s actual and expected wins is the seventh-biggest gap through eight games for any team since 1960. And 2020’s second-luckiest team by Pythagorean record is only fractionally behind: The Buffalo Bills rank ninth on the post-1960 list through nine games, with a gap of 2.3 wins between their actual 7-2 record and expected 4.7-win mark.
Who’s winning more than their margins predict?
2020 NFL teams by gap between actual wins and wins predicted by the Pythagorean expectation
Just because the Steelers and Bills are winning more than their scoring margins would predict doesn’t mean that they’re destined to collapse in the second half of the season. Indeed, our forecast thinks both teams are more than 90 percent likely to make the playoffs, ranking both among the top eight Super Bowl contenders by championship probability. (The Bills’ return to glory also remains one of the best storylines of 2020.) But it’s worth watching them for signs of regression after they won more than their share of close games in the first half of the schedule.
Conversely, we might also expect better things going forward from the New York Jets, Atlanta Falcons, Jacksonville Jaguars and Los Angeles Chargers. Yes, the Falcons and Chargers have developed an almost mythical ability to lose winnable games in their closing moments. According to the Pythagorean expectation, though, both Atlanta and L.A. “should” be around .500 but are running multiple wins behind that pace.
Neither the Chargers nor the Falcons are quite as historic an outlier as the Steelers or Bills are at the other end of the spectrum. Through nine games, Atlanta has the 131st-largest shortfall between actual and predicted record since 1960, and Los Angeles ranks 45th through eight. There are also limits to how much we can feel like these teams have been wronged by the football gods. The Chargers and Falcons each rank no higher than 20th leaguewide in schedule-adjusted expected points added (EPA) per game, so it’s not like their mediocre records are hiding great teams underneath.
But individually speaking, Chargers quarterback Justin Herbert is rising in the ranks of the historically unlucky, particularly after throwing what appeared to be a game-winning touchdown pass against the Raiders on Sunday — only to have it overturned on replay after time had expired.
Just seven starts into his NFL career, Herbert is off to a pretty incredible rookie season, with a passer rating of 104.7 (ninth-best in the league) and more than 300 passing yards per game. According to our QB Elo metric, Herbert is averaging a performance 85 Elo points better than league average across his seven starts, which ranks sixth in the NFL.3 Normally, we would associate that level of QB play with team success. And if we run a logistic regression between QB Elo over average and whether the QB’s team won a given game (holding other factors, such as the opposing QB’s performance, constant) — essentially creating an updated, Elo-based version of what I used to call the “Rivers Index,”4 — we would expect Herbert to have 4.8 wins as a starter in his young career.
Instead, at 1-6, Herbert has the biggest shortfall (-3.8 wins) of any QB in football this season according to the Rivers Index 2.0, easily topping Houston’s Deshaun Watson (-2.8 wins) for the No. 1 spot. In fact, only five QBs since 1960 have produced a bigger Rivers Index shortfall over their team’s first eight games of a season:
Justin Herbert should be winning more ballgames
NFL quarterbacks arranged by an updated version of the Rivers Index — the biggest shortfalls between actual and expected wins (based on QB stats) — through a season’s first eight games, 1960-2020
|Quarterback||Season||Team||Starts||Avg. Game Value*||Actual||Predicted||Rivers Index|
This franchise is no stranger to doing less with more. Secret Base’s Jon Bois and Alex Rubenstein did a great deep dive into the 2010 Chargers, who somehow missed the playoffs despite leading the league in both total yards gained and fewest yards allowed. (Their special teams unit was historically horrible, which helps explain the disconnect.) This Chargers team also has a notably bad special-teams corps, so in that sense history is repeating itself, and the rest has come down to back-breaking late penalties and other painful items on the “how to lose close games” laundry list.
[Related: The Bills Haven’t Been AFC East Champs In Josh Allen’s Lifetime. Odds Are, That Changes This Year.]
Just like the Steelers and Bills are under the microscope for backsliding down the stretch, we should keep an eye on the Chargers and Falcons to eventually start winning those toss-up contests late in the season. Or, who knows? Maybe they can just keep finding new ways to stay at the bottom of the Pythagorean disappointment rankings.
FiveThirtyEight’s NFL Elo ratings
How each team ranks through Week 9 of the 2020 season, according to our quarterback-adjusted predictions
|Chance To …|
|Rk||Team||Starting QB||QB Rk*||Elo Rating||Proj. Record||Make Playoffs||Win Div.||Win SB|
Looking ahead: Last time around, Elo’s game of the week — Packers versus 49ers — didn’t really end up being competitive at all. (Whoops!) This week’s top Elo game — Rams versus Seahawks — should be better, however, since both starting quarterbacks are at full strength (and neither club is fighting a COVID-19 outbreak). Both teams are in solid playoff position, though each lost in its last outing, calling into question their status as true contenders: Los Angeles was dealt a double-digit loss by the Dolphins’ dominating defense, while Seattle’s own D gave up 44 points at home to the Bills. Rams QB Jared Goff will try to exploit that weakness and break out of his recent slump, though the more compelling matchup involves the Rams defense (No. 4 in EPA) versus Russell Wilson and the Seahawks offense (also No. 4 in EPA). Seattle is the better team according to Elo, but with L.A. at home, we see a true 50-50 toss-up. Elo’s spread: Pick-’em
Check out our latest NFL predictions.