For those who are new to the feature, here's the deal: every week, I break down a topic related to regression to the mean. Some weeks, I'll explain what it is, how it works, why you hear so much about it, and how you can harness its power for yourself. In other weeks, I'll give 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 I'm looking at receivers and Justin Jefferson is one of the top performers in my sample, then Justin Jefferson goes into Group A, and may the fantasy gods show mercy on my predictions.

And then because predictions are meaningless without accountability, I track and report my results. Here's last year's season-ending recap, which covered the outcome of every prediction made in our seven-year history, giving our top-line record (41-13, a 76% hit rate) and lessons learned along the way.

Our Year to Date

Sometimes, I use this column to explain the concept of regression to the mean. In Week 2, I discussed what it is and what this column's primary goals would be. In Week 3, I explained how we could use regression to predict changes in future performance-- who would improve, who would decline-- without knowing anything about the players themselves. In Week 7, I explained why large samples are our biggest asset when attempting to benefit from regression. 

In Week 9, I gave a quick trick for evaluating whether unfamiliar statistics are likely stable or unstable. In Week 11, I explained the difference between regression and the gambler's fallacy, or the idea that players are "due" to perform a certain way. And in Week 12, I showed how understanding regression can allow us to predict the past as easily as the future.

Sometimes, I point out broad trends. In Week 5, I shared twelve years worth of data demonstrating that preseason ADP held as much predictive power as performance to date through the first four weeks of the season.

Other times, I use this column to make specific predictions. In Week 4, I explained that touchdowns tend to follow yards and predicted that the players with the highest yard-to-touchdown ratios would begin outscoring the players with the lowest. In Week 6, I explained that yards per carry was a step away from a random number generator and predicted the players with the lowest averages would outrush those with the highest going forward.

In Week 8, I broke down how teams with unusual home/road splits usually performed going forward and predicted the Cowboys would be better at home than on the road for the rest of the season. In Week 10, I explained why interceptions varied so much from sample to sample and predicted that the teams throwing the fewest interceptions would pass the teams throwing the most.

The Scorecard

The prediction is only halfway done, but it's looking very unlikely the Cowboys will salvage things at this point and finish the year better at home than on the road. If the prediction is to have any shot, a blowout win against the Giants on Thanksgiving would be a good place to start.

Our interception prediction is faring much better. Again, a mistake in my math led to making the prediction tougher than I intended, but I needn't have worried; Group B isn't just throwing fewer total interceptions; they're even averaging fewer per game.

Most Players Regress. Rookies Progress.

We've talked this season about how everyone regresses to the mean, but everyone's mean is different. Thinking of player performance as random fluctuations around a fixed "true mean" is useful. But it's not maximally accurate.

Just as means can vary from player to player, they can also change over time. Randall Cunningham was one of the most prolific rushing quarterbacks in history. In his 20s, he averaged 41.1 rushing yards per game at 7.0 yards per carry. This was his "true mean".

In his 30s, he averaged 14.7 rushing yards per game at 4.7 yards per carry. This was also his true mean. Quarterback rushing tends to age much like running back rushing, with even the most prolific runners finding themselves running less frequently and less successfully.

How do you estimate a mean that's a moving target? There are techniques one could use. One could use a so-called "moving average", which places extra weight on more recent data, making it more sensitive to trends. Or one could develop a good working knowledge of general patterns (such as quarterback rushing declining around age 30) and update expectations accordingly.

Today we're going to do the latter. Players tend to score the same number of fantasy points in the first half and the second half of the season. But I have a working knowledge of one class of players whose performance tends to trend up over time: rookies.

Instead of level performance, rookies (as a class) have a strong tendency to score more points later in the season. There are plenty of good reasons for this -- increased trust from coaches and teammates, expanded roles from teams working them in slowly, gaining a better feel for the game as they gain more experience. Those reasons don't change the fact that instead of regressing towards their "true mean", rookies appear to regress away from it late in the year.

Except they don't really. They aren't the lone exception to the rule that everyone regresses toward their true mean; their true mean simply rises over time, giving the appearance of anti-regression. (Good fantasy managers can take advantage of this fact to draft extra rookies in September so that their team tends to get better late in the season when the points matter more.)

Given the knowledge that their mean rises throughout the season, we can make a prediction. 

Sixteen wide receivers were drafted in the first three rounds of the NFL draft: Marvin Harrison Jr., Malik Nabers, Rome Odunze, Brian Thomas Jr., Xavier Worthy, Ricky Pearsall, Xavier Legette, Keon Coleman, Ladd McConkey, Ja'Lynn Polk, Adonai Mitchell, Malachi Corley, Jermaine Burton, Roman Wilson, Jalen McMillan, and Luke McCaffrey. Wilson is on injured reserve, which leaves fifteen.

Collectively, that group has played 135 games and scored 1111.2 PPR fantasy points for an average of 8.23 points per game. If their mean is really on the rise, the group should have no trouble beating that average over the next four weeks. Additionally, I predict that this won't be the function of a few strong performers; at least 60% of those receivers (9 of 15) will average more points per game over the next four weeks than they did over the first twelve.

It's worth noting that they'll have to overcome headwinds to do so. I mentioned that most players perform about as well in the first half and second half of the season, but that's not entirely true; as the weather turns colder, NFL passing declines a bit, which means most receiving cohorts see a slight dip in November and December.

Receivers from the 2018, 2019, 2020, 2021, 2022, and 2023 classes will probably score slightly less over the next month. The 2024 class, though? I think they're going to see an improvement in production. We'll have to check back in four weeks to find out if I was right.