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 yards per target, and Antonio Brown is one of the high outliers in yards per target, then Antonio Brown 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 all my predictions from last year and how they fared.
THE SCORECARD
In Week 2, I laid out our guiding principles for Regression Alert. No specific prediction was made.
In Week 3, I discussed why yards per carry is the least useful statistic and predicted that the rushers with the lowest yard-per-carry average to that point would outrush the rushers with the highest yard-per-carry average going forward.
In Week 4, I explained why touchdowns follow yards, (but yards don't follow back), and predicted that the players with the fewest touchdowns per yard gained would outscore the players with the most touchdowns per yard gained going forward.
In Week 5, I talked about how preseason expectations still held as much predictive power as performance through four weeks. No specific prediction was made.
In Week 6, I looked at how much yards per target is influenced by a receiver's role, how some receivers' per-target averages deviated from what we'd expect according to their role, and predicted that the receivers with the fewest yards per target would gain more receiving yards than the receivers with the most yards per target going forward.
In Week 7, I demonstrated how randomness could reign over smaller samples, but regression dominates over larger ones. No specific prediction was made.
In Week 8, I discussed how even something like average career length could be largely determined by regression-prone fluctuations in incoming talent. No specific prediction was made.
In Week 9, I looked at running backs scoring touchdowns at an unsustainable rate and posited that even Todd Gurley must return to earth.
In Week 10, I delved into the purpose of regression alert and the proper takeaways. No specific prediction was made.
In Week 11, I explained an easy way to find statistics that were more prone to regression and picked on yards per carry one more time.
In Week 12, I went into the difference between regression to the mean, (the idea that production will probably improve or decline going forward), and the gambler's fallacy, (the idea that production is "due" to improve or decline going forward). No specific prediction was made.
Statistic For Regression
|
Performance Before Prediction
|
Performance Since Prediction
|
Weeks Remaining
|
Yards per Carry
|
Group A had 24% more rushing yards per game
|
Group B has 4% more rushing yards per game
|
SUCCESS!
|
Yards:Touchdown Ratio
|
Group A had 28% more fantasy points per game
|
Group B has 23% more fantasy points per game
|
SUCCESS!
|
Yards per Target
|
Group A had 16% more receiving yards per game
|
Group A has 13% more receiving yards per game
|
Failure
|
Yards:Touchdown Ratio
|
Group A had 26% more fantasy points per game
|
Group B has 4% more fantasy points per game
|
SUCCESS!
|
Yards per Carry
|
Group A had 9% more rushing yards per game
|
Group B has 14% more rushing yards per game
|
2
|
This week, we close the book once again on Yard to Touchdown ratio, (the "Todd Gurley prediction"), after Group B staged a dramatic comeback to steal a last-minute victory.
Now that the prediction is closed and it doesn't come off as self-serving rationalization, I can admit that this was easily the prediction I was most nervous about. Classically, the goal is to pick a Group B that averages more yards than Group A, but fewer touchdowns. And that's what I did, originally. But adding Todd Gurley to Group A at the end left those backs not just with (a lot) more touchdowns, but also with slightly more yards, as well. And then after the massive washout in the first week I figured Group B was doomed.
But Christian McCaffrey, Ezekiel Elliott, and Saquon Barkley ultimately averaged over 150 yards and nearly 1.5 touchdowns per game, (offsetting an underperforming trio of Adrian Peterson, Tarik Cohen, and Joe Mixon), Gurley fell from incandescent to merely great, (he ranked 5th out of the 8 Group A backs in fantasy points per game), and Group B managed to rally to victory.
There are a lot of amazing stats from the last four weeks, but there's one in particular I wanted to highlight. Remember, the groups were created based on the frequency with which running backs were scoring touchdowns. The backs in Group A scored nearly twice as many touchdowns per game as the backs in Group B through eight weeks. In the four weeks since, the Group B backs not only scored more touchdowns per game, they actually had a bigger advantage in touchdowns scored (12%) than they did in yards gained (2%)! Touchdowns are vitally important for fantasy football, but they're also incredibly random and past touchdown production is not the sort of thing you want to be pinning your hopes to going forward.
Finally, a few words about the most recent yards per carry prediction. Over the last two weeks, Group B isn't just averaging more carries per game than Group A, as we expected. They're also averaging more yards per carry.
Picking on Quarterbacks
I know I haven't made any quarterback-specific predictions so far this year. In large part, that's because "efficiency stats" like yards per attempt are much more stable for quarterbacks than the equivalent stats for running backs (yards per carry) or receivers (yards per target). My favorite regression target for quarterbacks is yard to touchdown ratio, just like at running back and wide receiver... but this year there are surprisingly few outliers available to create a competitive Group B. (The "worst" quarterback this year still averages just 221 yards per touchdown. From 2013-2017 we saw 33 quarterbacks end the season with higher ratios, an average of 6.6 per year.)
There is one quarterback statistic that I loathe with a fire normally only reserved for yards per carry, though: interception rate.
Fellow Footballguy Danny Tuccitto once calculated how many attempts a player needed before that player's "per-attempt" statistics were 50% the result of the player's underlying skill and 50% the result of random chance. Yards per carry, the example I keep bringing up of a statistic that is almost entirely noise, took 1,978 carries to stabilize. Interception% took 1,681 pass attempts. (Yards per attempt, by contrast, stabilizes in just 396 attempts.)
More importantly, that's not 1,681 career attempts, that's 1,681 attempts in the same system and with (largely) the same supporting cast. So while Jameis Winston has 1,746 career pass attempts, he's nowhere near the point where we can say that his career interception percentage is mostly representative of his true level of play. Which seems like an important point to make given that Jameis Winston keeps getting benched for throwing interceptions.
So let's highlight how silly it is to use past interception rates to forecast future interception rates. Setting aside any rookies, (because rookies are rookies), there are 27 players who have thrown for at least 1,500 yards so far this year. Andy Dalton, Joe Flacco, and Alex Smith are injured, and Blake Bortles just got benched, leaving us with 23 remaining candidates.
Of those 23 players, 14 have thrown eight or fewer interceptions. These players are: Derek Carr, Cam Newton, Eli Manning, Tom Brady, Kirk Cousins, Marcus Mariota, Carson Wentz, Philip Rivers, Jared Goff, Russell Wilson, Dak Prescott, Matt Ryan, Drew Brees, and Aaron Rodgers. This is our Group A. Quarterbacks in Group A have collectively thrown 78 interceptions on 5,136 attempts for an interception rate of 1.52%.
On the other hand, Deshaun Watson, Mitchell Trubisky, Matthew Stafford, Patrick Mahomes II, Case Keenum, Andrew Luck, Jameis Winston, Ben Roethlisberger, and Ryan Fitzpatrick have all thrown nine or more interceptions. This is our Group B. Group B quarterbacks have combined for 94 interceptions on 3,198 attempts, an interception rate of 2.94%, basically double that of Group A.
Despite Group A throwing 17% fewer interceptions to this point of the season, I predict that because Group B has fewer quarterbacks throwing fewer attempts, they'll finish out the next four weeks also throwing fewer interceptions.