Regression Alert: Week 8 - Footballguys

Anything that is subject to randomness is subject to regression, (including some surprising things you might not think of at first)

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.

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 6% more receiving yards per game
2

Our second specific prediction of the week is now on the books, and regression once again dominated. At the time of my prediction, Group A had scored 36 touchdowns in 39 games and Group B had scored 4 touchdowns in 36 games. Since my prediction, Group A scored 19 touchdowns in 42 games, while Group B scored 22 touchdowns in 43 games. That's right, the "low-touchdown" group reached the end zone more frequently than the "high-touchdown" group! Given time, the touchdowns always follow the yards.

As for the yards per target crowd, we continue to see some small regression, but not quite enough to completely flip the groups. As I said last week, though, time is regression's greatest ally, so we'll see how they fare over the final two weeks of the sample.


When Trends Are Noise

Quarterbacks are playing longer than ever these days. Last year, Tom Brady set the record for oldest player ever named MVP by the Associated Press, breaking a record that had previously been set by... Peyton Manning, just four years earlier. The year before, he became the second-oldest quarterback to ever win a Super Bowl, behind only Peyton Manning in 2015. This year, Drew Brees is an early front-runner to be named MVP. He'd pass Peyton for the 2nd-oldest in history. Philip Rivers is also in contention, and if he wins he'll be the 3rd-oldest AP MVP.

Why are quarterbacks playing so long, though? Is it because advances in training, nutrition, and medicine are extending careers? I'm sure that's part of it, but if that's the entire story it becomes hard to explain why running backs aren't seeing their careers similarly extended. With all due respect to Adrian Peterson and Frank Gore, a 30-year-old running back rushed for 1,000 yards 21 times from 2000-2009. They only did it six times in the eight years since.

Is it because the league is shifting toward the pass and the NFL is instituting more and more rules designed to protect quarterbacks? That undoubtedly plays a role, but if that's the entire story we're left to explain how wide receivers haven't seen a similar increase in longevity thanks to more roster spots being devoted to the position and protections being granted to defenseless receivers. With apologies to Larry Fitzgerald, a receiver topped 1,000 yards at age 34 or older 20 times between 2000 and 2009. They've only managed it four times since.

In fact, there's no other position where careers are getting longer like they are at quarterback. In the last decade, eight different offensive linemen have started at least half a season at age 36 or older. Six different players did it in the year 2000 alone. There were seven different double-digit sack seasons by a 36-year-old player between 1997 and 2000. There has been one in the 17 years since. Even kickers aren't seeing any major improvements. From 2000-2009, the league averaged three kickers and punters per year over the age of 40. From 2010-2017, it averaged 2.5. (Old placekickers were slightly up, but old punters were way down.)

So the question isn't "why are quarterbacks playing longer than ever?" The real question here is "why are only quarterbacks playing longer than ever?" And that's a question that I think has a very simple answer: the 2000-2005 NFL drafts were amazing for quarterbacks.

In that six-year span, the league gained sure-fire first-ballot Hall of Famers Tom Brady, Aaron Rodgers, and Drew Brees, as well as likely-to-borderline Hall of Famers Ben Roethlisberger, Philip Rivers, Eli Manning, and Tony Romo, and multiple pro-bowlers Alex Smith, Carson Palmer, Michael Vick, Matt Schaub, and Marc Bulger.

Compare that to the six years immediately prior, (1994-1999). Those drafts gave us Peyton Manning, fringe Hall of Famers Donovan McNabb and Steve McNair, multiple pro-bowlers Matt Hasselbeck, Daunte Culpepper, and Kerry Collins, and... that's it. Now look at the six years immediately after, (2006-2011). Those drafts produced a pair of players on a fringe Hall-of-Fame trajectory in Cam Newton and Matt Ryan, as well as long-time starters Matthew Stafford, Jay Cutler, Joe Flacco, and Andy Dalton, plus guy-who-actually-made-two-pro-bowls Vince Young. (Seriously, that happened.)

In fact— and this is the really amazing part— there are more active quarterbacks in the league today who were drafted between 2000 and 2005 (12 total) than were drafted between 2006 and 2011 (11 total). It's not that quarterbacks are playing longer. It's that we had a truly spectacular influx of quarterback talent and now it's getting old, and not enough young quarterback talent came in afterward to take its place.

In fact, so many fluctuations in average age are simply the result of fluctuations in incoming talent. Remember that stat about how there were seven double-digit sack seasons by a player 36 or older between 1997 and 2000? There were two from Hall of Famer Reggie White, two from Hall of Famer Chris Doleman, two from Hall of Famer Kevin Greene, and one from Hall of Famer Bruce Smith. It's not a mystery why pass rushers were suddenly playing longer— it's because a handful of amazing pass rushers entered the league in a short span and they all got old at the same time.

But incoming talent is largely random, which means it's subject to regression. Some years are good years for quarterbacks, other years are bad ones. Sometimes you get a few great years in a row. Sometimes you go long spans without any great years. Random is random.

The problem is when these random fluctuations get mistaken for structural trends. A lot of people assume today that quarterbacks playing into their late 30s is the new normal. It's not! Hall of Fame quarterbacks playing into their late 30s is the old normal. Sammy Baugh did it in the '50s; Bobby Layne and Y.A. Tittle did it in the '60s; Johnny Unitas, Sonny Jurgensen, and Fran Tarkenton did it in the '70s; Moon, Montana, Elway, Marino, and Young did it in the '90s; Favre and Warner did it in the '00s; and now Manning, Brady, Brees, Rivers, Roethlisberger, Rodgers, and company are doing it in the '10s. (Why did no one do it in the '80s? Again, fluctuations in incoming talent. The only Hall of Fame quarterbacks who entered the league between 1970 and 1978 were Terry Bradshaw and Dan Fouts, whose careers wound down in their mid-30s.)

This idea that a lot of "trends" are nothing more than ebbs and flows in incoming talent has especially big implications in dynasty leagues. I started playing dynasty in 2007, and the overwhelming consensus at the time was that you had to build your team around running backs, because we had just passed a massive boom in running back talent, (LaDainian Tomlinson, Priest Holmes, Ricky Williams, Shaun Alexander, Clinton Portis, Marshall Faulk, et al.) This prevailing wisdom was only reinforced in the next few years thanks to the 2008 draft class, arguably the best overall collection of running back talent in modern NFL history.

By 2015, this conventional wisdom was completely flipped on its head. The true path to success, everyone realized at the same time, was actually to build around wide receivers. The NFL was a passing league, now! It didn't even care about running backs anymore. After all, the 2011 draft didn't see a running back come off the board until the 28th pick, the latest teams had ever waited in the common draft era. And then in 2013, for the first time, the entire first round went by without a running back being selected. And in 2014 it happened again!

Meanwhile, star wide receivers were entering the NFL at a ridiculous pace. 2010 brought Antonio Brown, Demaryius Thomas, and Dez Bryant. 2011 gave us A.J. Green, Julio Jones, and Randall Cobb. 2012 had T.Y. Hilton and Alshon Jeffery; 2013 had DeAndre Hopkins and Keenan Allen. And then, the pièce de résistance, the 2014 class, arguably the best collection of incoming wide receivers the league has ever seen: Odell Beckham Jr, Mike Evans, Brandin Cooks, Jarvis Landry, Sammy Watkins, Davante Adams, Allen Robinson, Adam Thielen... even Kelvin Benjamin, Jordan Matthews, Donte Moncrief, John Brown, Marqise Lee, Quincy Enunwa, Martavis Bryant, and Paul Richardson Jr had fantasy value.

But you probably see where this is going already; the 2009-2014 drafts weren't indicative of anything larger than the fact that 2009-2014 happened to be incredible years for incoming receivers and terrible years for incoming backs. The five years since, however, have given us 8 pro bowl running backs (Todd Gurley, Ezekiel Elliott, Melvin Gordon III, David Johnson, Kareem Hunt, Alvin Kamara, Jay Ajayi, and Jordan Howard, with Saquon Barkley and James Conner decent bets to add to that total this year), vs. just 3 pro bowl wide receivers (Amari Cooper, Tyreek Hill, and Michael Thomas; Tyler Lockett and Pharoh Cooper also made the pro bowl for their work as returners).

Here's a chart that helps visualize the idea that incoming talent is random and everything that's random is prone to regression. Here's the average age of the top 24 fantasy quarterbacks, top 36 running backs, top 36 receivers, and top 24 tight ends. This age is weighted by fantasy production, so the #1 player contributes more to the average than the #24. In theory, if no new talent came in we should expect the average age to tick up by ~1 per year as everyone became a year older. So when the age increases, that suggests veterans are holding off young challengers, while a decreasing average age indicates a youth movement at the position.

I've shaded the cells to make trends easier to see; RED means the position as a whole is younger than average, while BLUE means that the position is older.

Looking at the chart, you can see the fluctuations in talent working their way across the league. The 2000-2005 quarterbacks took over in the late 2000s and have been slowly aging since. The 2008 running back class had an immediate impact across the league, as have the recent classes. Average receiver age plummeted nearly immediately after the historic 2014 class. The tight end position has been aging alongside Rob Gronkowski and Jimmy Graham since their breakout in 2011.

But it's important to remember that these trends aren't trends at all. They're descriptions of a random process. That quarterback list isn't going to keep getting bluer and bluer over time; the 2000-2005 quarterbacks will eventually retire, a new crop of young stars will take their place, and the chart will tilt back into the red. The 2019-2021 classes are unlikely to produce as many star running backs as the 2016-2018 classes did, so the position will likely start to age alongside Gurley, Elliott, Gordon, Hunt, Kamara, and Barkley. We'll doubtless see more stellar young receivers start to enter the league at some point to help offset the aging of 2010-2014's stars. We're already seeing new talent emerge at tight end.

This isn't to say that the NFL never has any trends. The league really is trending toward the pass, even if talent disparities might have exacerbated the process in recent years. Kickers are getting more accurate every year, (even if the salience of a few missed extra points suggests otherwise). But for the most part, when faced with an apparent trend, it's best to remember the old saying: it's not the X's and O's, it's the Jimmys and Joes.


More articles from Adam Harstad

See all

More articles on: Forecast

See all

More articles on: History

See all

More articles on: QB

See all