Unlock More Content Like This With A
Footballguys Premium Subscription
"Footballguys is the best premium
fantasy football only site on the planet."
Matthew Berry, NBC Sports EDGE
Welcome to Regression Alert, your weekly guide to using regression to predict the future with uncanny accuracy.
For those who are new to the feature, here's the deal: every week, I dive into the topic of regression to the mean. Sometimes I'll explain what it really is, why you hear so much about it, and how you can harness its power for yourself. Sometimes I'll give some practical examples of regression at work.
In weeks where I'm giving practical examples, I will select a metric to focus on. I'll rank all players in the league according to that metric, and separate the top players into Group A and the bottom players into Group B. I will verify that the players in Group A have outscored the players in Group B to that point in the season. And then I will predict that, by the magic of regression, Group B will outscore Group A going forward.
Crucially, I don't get to pick my samples (other than choosing which metric to focus on). If the metric I'm focusing on is touchdown rate, and Christian McCaffrey is one of the high outliers in touchdown rate, then Christian McCaffrey goes into Group A, and may the fantasy gods show mercy on my predictions.
Most importantly, because predictions mean nothing without accountability, I track the results of my predictions over the course of the season and highlight when they prove correct and also when they prove incorrect. Here's a list of my predictions from 2019 and their final results, here's the list from 2018, and here's the list from 2017.
THE SCORECARD
In Week 2, I opened with a primer on what regression to the mean was, how it worked, and how we would use it to our advantage. No specific prediction was made.
In Week 3, I dove into the reasons why yards per carry is almost entirely noise, shared some research to that effect, and predicted that the sample of backs with lots of carries but a poor per-carry average would outrush the sample with fewer carries but more yards per carry.
In Week 4, I talked about how the ability to convert yards into touchdowns was most certainly a skill, but it was a skill that operated within a fairly narrow and clearly-defined range, and any values outside of that range were probably just random noise and therefore due to regress. I predicted that high-yardage, low-touchdown receivers would outscore low-yardage, high-touchdown receivers going forward.
In Week 5, I talked about how historical patterns suggested we had just reached the informational tipping point, the time when performance to this point in the season carried as much predictive power as ADP. In general, I predicted that players whose early performance differed substantially from their ADP would tend to move toward a point between their early performance and their draft position, but no specific prediction was made.
In Week 6, I talked about simple ways to tell whether a statistic was especially likely to regress or not. No specific prediction was made.
In Week 7, I speculated that kickers were people, too, and lamented the fact that I'd never discussed them in this column before. To remedy that, I identified teams that were scoring "too many" field goals relative to touchdowns and "too many" touchdowns relative to field goals and predicted that scoring mix would regress and kickers from the latter teams would outperform kickers from the former going forward.
In Week 8, I noted that more-granular measures of performance tended to be more stable than less-granular measures and predicted that teams with a great point differential would win more games going forward than teams with an identical record, but substantially worse point differential.
Statistic for regression | Performance before prediction | Performance since prediction | Weeks remaining |
---|---|---|---|
Yards per Carry | Group A had 3% more rushing yards per game | Group B has 36% more rushing yards per game | Success! |
Yard to Touchdown Ratio | Group A averaged 2% more fantasy points per game | Group B averages 40% more fantasy points per game | Success! |
TD to FG ratio | Group A averaged 20% more points per game | Group B averages 24% more points per game | 2 |
Wins vs. Points | Both groups had an identical win% | Group B has a 17% higher win% | 3 |
Week 8 was an unusually compressed week for kicker scoring (half of the active kickers finished with between 4 and 7 points), which naturally tightened up the race between Group A and Group B. But Group B maintains a solid edge with two weeks to go.
The way I originally framed the wins vs. point differential prediction meant that we were expecting Group A to outperform Group B going forward. For the sake of consistency (to make sure we're always rooting for Group B in these comparisons), I've decided to just switch the labels. The groups are the same, but we'll call the high-point-differential teams "Group B" instead of "Group A" so the chart doesn't get confusing.
Anyway, our Group B (the high-differential teams) went 3-2, while our Group A (the low-differential teams) went 3-4, which means Group B's win% is 17% higher through one week. (Group B also maintained a point-differential advantage of 4 points per game.)
Regression and Dynasty
Regression to the mean is a pretty straightforward affair in redraft leagues, where all that matters is how many points a player will score you over the next few weeks or months. Given two guys who are comparably productive to this point, you want the guy who will score more points going forward.
I've outlined some simple rules you can use to try to help identify which players will sustain their production and which won't. Granular measures (yards, point differential) tend to be more reliable indicators than blunter measures (touchdowns, wins). Anything that is heavily influenced by outliers (such as yards per carry, which is dominated by long runs) tends to be less reliable than something that isn't (success rate).
Things that are wholly within a team's control (pass to run ratio, targets) tend to be more reliable than things that depend on the opponent's performance, too (yards). Additionally, if we're unsure about a metric we can simply look at the leaderboard to see if it's dominated by good players (in which case it's probably measuring something about player talent) or whether it's a mixed bag (in which case it's probably dominated by randomness).
Sometimes things get tricky. Yards per target is the classic example; like yards per carry, it is dominated by outlier plays and thus prone to regression. But unlike yards per carry, the top of the leaderboard is dominated by great players, which means that talent is a big factor in performance, and since talent is sticky yards per target is sticky, too. That's why no longer include yards per target in Regression Alert.
But while it's easy to say what to do with players who are destined to regress in redraft (sell them for more than they're worth), the answer is much trickier in dynasty. I often write about how no one is immune to the forces of regression. For instance, in 2017 I wrote that Alvin Kamara was averaging a touchdown for every 15.5 touches, which would be a ridiculous historical outlier. Since then, Kamara has scored one touchdown for every 21.6 touches, which is a very high rate, but in the same range as players like Shaun Alexander, Priest Holmes, and LaDainian Tomlinson.
In 2018, I noted that Todd Gurley had scored 15 touchdowns in 8 games, which was well beyond the limits of sustainability. Since then he's scored 28 touchdowns in 29 games, about half the rate he was previously scoring. Early in Rob Gronkowski's career, I speculated that he might be dominant enough in the red zone for his yard-to-touchdown ratio to stay south of 100. It didn't.
In 2015, I wrote that players struggle to score more than 1 touchdown per 100 yards, but Gronk might be an exception.
— Adam Harstad (@AdamHarstad) November 21, 2018
Here's his career Yard:TD ratio after every season.
2010: 54.6
2011: 67.0
2012: 68.3
2013: 75.7
2014: 79.7
2015: 84.2
2016: 88.4
2017: 93.3
2018 (so far): 97.8
Just realized: last week’s game has put Rob Gronkowski over 100 receiving yards per receiving touchdown for the first time in his entire career. (His first career catch was a 1-yard touchdown, natch.)
— Adam Harstad (@AdamHarstad) October 2, 2020
Last year I noted that Lamar Jackson's 78 yards per game rushing was more than 50% higher than Michael Vick's career average and almost double Randall Cunningham's. This year, Jackson averages 59 yards per game. I've talked about how Patrick Mahomes II' 8.6% touchdown rate in 2017 was unsustainable (it's 6.1% since). I've said the same about Deshaun Watson's 9.3% rookie touchdown rate (it's 5.4% since). I could go on about Kyler Murray averaging a rushing touchdown per game (unsustainable), Dak Prescott's throwing for 370 yards per game (unsustainable), Russell Wilson throwing a touchdown on over 10% of his pass attempts (unsustainable), or Joe Burrow averaging 41 pass attempts per game (unsustainable).
And I could unload on Justin Herbert, the inspiration for this week's column. Herbert has scored 14 touchdowns in his last four games and is putting together the kind of rookie season that draws comparisons to Dan Marino, the best rookie quarterback in NFL history.
First 5 NFL starts
— Adam Harstad (@AdamHarstad) October 26, 2020
Dan Marino - 12 TDs, 3 INTs, 8.17 YPA, 0.37 yards lost to sacks per dropback (4.6% sack rate), 105.6 passer rating, 8.12 ANY/A.
Justin Herbert - 12 TDs, 3 INTs, 8.38 YPA, 0.35 yards lost to sacks per dropback (5.2% sack rate), 108.1 passer rating, 8.15 ANY/A.
But here's the thing. Alvin Kamara's touchdown for every 21 touches is still incredible. So is Todd Gurley's 28 touchdowns in his last 29 games. Lamar Jackson's 59 rushing yards per game would still be the best mark in history by a quarterback. Patrick Mahomes II's touchdown rate since 2017 would still be the best since the merger. Russell Wilson's touchdown rate before 2020 would still rank 3rd. Deshaun Watson's touchdown rate since his rookie year would rank in the top 10, right between Tom Brady and Drew Brees.
Just because a player is guaranteed to regress doesn't mean that player isn't still amazing. Really, the fact that a player is putting up such an unsustainable performance is evidence of his greatness. Patrick Mahomes II, Lamar Jackson, Kyler Murray, Deshaun Watson, Russell Wilson, Dak Prescott, Joe Burrow, and Justin Herbert are my top eight dynasty quarterbacks.
Now, if someone wants to trade for Justin Herbert because they think he's the next Dan Marino, and they're willing to pay an exorbitant price to match that belief, then sure, you should sell him. But by and large, long-term success boils down to loading up on talented players.
No one is immune to the forces of regression. No one. And yet, in the long run, it helps to load up on players who are bound to regress, because even after they do they're usually still among the best in the NFL.