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 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 Jaxon Smith-Njigba is one of the top performers in my sample, then Jaxon Smith-Njigba goes into Group A, and may the fantasy gods have 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 nine-year history, giving our top-line record (51-17, a 75% hit rate) and lessons learned along the way.

Our Year to Date

In Week 2, I laid out the four primary goals of Regression Alert. The first was to persuade you that regression to the mean is a real, actionable force in fantasy football. The core of that effort is making specific, concrete predictions based on nothing but regression to the mean and tracking the results over time. Here's a list of every prediction we have made so far:

1. I predicted that high-volume, low-ypc running backs would outrush lower-volume, high-ypc backs over the next four weeks (Week 4)

The Scorecard

Not a great start, but such is the nature of variance. Our high-ypc backs managed to increase not only their volume (Kyle Monangai, in particular, had 30 carries in the first three weeks then 30 more in Week 4); they also increased their ypc average (from 5.56 over the first three weeks to 6.00 in Week 4).

Weird stuff happens in single-week samples, which is why we run our predictions for a month.

Revisiting Preseason Expectations

In October of 2013, I wondered just how many weeks it took before the early-season performance wasn't a fluke anymore. In "Revisiting Preseason Expectations", I looked back at the 2012 season and compared how well production in a player's first four games predicted production in his last 12 games. And since that number was meaningless without context, I compared how his preseason ADP predicted production in his last 12 games.

I didn't realize at the time that this would turn Week 5 into my own personal Groundhog Day. 

It was a fortuitous time to ask that question, as it turns out, because I discovered that after four weeks in 2012, preseason ADP still predicted performance going forward better than early-season production did.

This is the kind of surprising result that I love, but sometimes results are surprising because they're flukes. So, in October of 2014, I revisited "Revisiting Preseason Expectations". This time, I found that in the 2013 season, preseason ADP and week 1-4 performance held essentially identical predictive power for the rest of the season.

With two different results in two years, it was time for a tiebreaker. In October of 2015, I revisited my revisitation of "Revisiting Preseason Expectations". This time, I found that early-season performance held a slight predictive edge over preseason ADP. Like a dog chewing on a bone, when October rolled around in 2016, I revisited my revisitation of the revisited "Revisiting Preseason Expectations". And again in October 2017.

Now fully a creature of habit, when October 2018 rolled around, I simply had to revisit my revisitation of the revisited revisited revisitation of "Revisiting Preseason Expectations". And then in October 2019, and October 2020, and October 2021, and October 2022, and October 2023, and October 2024, and October 2025, I... well, you get the idea.

And now, as you've probably guessed, it's time for an autumn tradition as sacred as turning off the lights and pretending I'm not home on October 31st. It's time for the fourteenth annual edition of "Revisiting Preseason Expectations"! (Or as I prefer to call it, "Revisiting Revisiting Revisiting Revisiting Revisiting Revisiting Revisiting Revisiting Revisiting Revisiting Revisiting Revisiting Revisiting Revisiting Preseason Expectations".)

Methodology

If you've read the previous pieces, you have a rough idea of how this works, but here's a quick rundown of the methodology. I have compiled a list of the top 24 quarterbacks, 36 running backs, 48 wide receivers, and 24 tight ends by 2025 preseason ADP.

From that list, I have removed any player who missed more than one of his team's first four games or more than two of his team's last thirteen games so that any fluctuations represent performance and not injury. As always, we're looking by team games rather than by week, so players with an early bye aren't skewing the comparisons (though this wasn't a factor last year, as all byes happened after Week 4).

I have always used PPR scoring for this exercise because that was easier for me to look up with the databases I had on hand a decade ago. For everyone who didn't miss significant time, I tracked where they ranked at their position over the first four games and over the final thirteen games. Finally, I've calculated the correlation between preseason ADP and stretch performance, as well as the correlation between early performance and stretch performance.

Correlation is a measure of how strongly one list resembles another list. The strongest possible correlation is 1, which is what you get when two lists are identical. The weakest possible correlation is 0, which is what you get when you compare one list of numbers to a second list that has no relationship whatsoever. (Correlations can go down to -1, which means the higher something ranks in one list, the lower it tends to rank in the other, but correlations near -1 also represent very strong relationships.)

So if guys who were drafted high in preseason tend to score a lot of points from weeks 5-18, and this tendency is strong, we'll see correlations closer to 1. If they don't tend to score more points, or they do, but the tendency is very weak, we'll see correlations closer to zero. The numbers themselves don't matter beyond "closer to 1 = more predictable".

For the sake of transparency, I'll post the raw data from last year. This largely isn't important; I'd recommend most readers skip down to the "Overall Conclusions" section below for the key takeaways.  

Quarterback

Running Back

Wide Receiver

Tight End

Overall Conclusions

We could cherry-pick individual names from those lists to argue for ADP or early-season performance. Ja'Marr Chase and Justin Jefferson were the Top 2 receivers by preseason ADP last year. They ranked 17th and 15th, respectively, through the first month of the season. Chase rebounded to finish 3rd the rest of the way; his early-season results were a mirage. Jefferson fell off further and ranked 27th down the stretch; his early-season results were a harbinger.

But going name by name won't get us anywhere quickly, so here is the data on correlation between ADP and stretch performance, early-season performance and stretch performance, and a simple average of both factors and stretch performance.

Note that I am only displaying values to three digits, so if some of the comparisons are off by a thousandth of a point or so, this is likely due to rounding.

Quarterback

Running Back

Wide Receiver

Tight End

All Positions

These correlations are as close as can be. Just five thousandths of a point separates ADP from early-season performance over the last decade.

I've run this study fourteen times now. In seven of them, preseason ADP was more predictive. In six, early-season results mattered more. And in one, the difference was 0.001—indistinguishable from zero.

The square of the correlation is said to represent how much of the variation in the second dataset is explained by the variation in the first. By that measure, preseason ADP explains 31.1% of the variation in rest-of-year production. Early-season performance explains 30.5%. (The average of the two factors beats either factor alone, explaining 38.7% of the variation.)

If your league had an annual tradition of drafting entirely new teams heading into Week 5, and one of your leaguemates had an annual tradition of drafting exclusively off of his or her pre-draft list, while another leaguemate had an annual tradition of drafting exclusively off of results to date... both leaguemates would probably have nearly identical records.

What About Sub-Samples?

Overall, preseason ADP and early-season performance are likely equally predictive of rest-of-year results. But what about in specific cases? I've investigated a lot of sub-samples over the last decade+ of running these numbers, and here are the results:

• Is early-season performance more predictive for tight ends?

In 2019, I noticed that early-season results beat preseason ADP for TEs in each of the last five seasons and speculated that the position might be the exception to the overall rule. Preseason ADP has beaten early-season performance in five of the seven seasons since, and in 2024 early-season performance was actually slightly negatively correlated with rest-of-year production (just the third negative correlation we've encountered in our sample).

Verdict: probably not.

• Is one or the other sample more predictive for late-round sleepers?

Looking *JUST* at the bottom 50% of the sample in terms of preseason ADP, here are the same correlations—
ADP: .436
Early points: .463
Avg of both: .516

Maybe a teeny edge to early performance? But that's so close I'd call it a wash.

— Adam Harstad (@AdamHarstad) September 10, 2023

Verdict: probably not.

• Is one factor more important for players who were early-season disappointments?

Out of concern I'm unknowingly conditioning on a collider, I repeated looking just at the bottom 50% of the sample in terms of Week 1-4 performance. Correlations—
ADP: .487
Early points: .455
Avg both: .526

— Adam Harstad (@AdamHarstad) September 10, 2023

Verdict: probably not.

• Is early-season performance more predictive for quarterbacks?

Last year, I noted that quarterback correlations had been trending toward early-season performance over time and it was now the lone position where early-season results actually outperformed the two-factor average. I mentioned this was probably a mirage, but promised to monitor it over time. The 2025 season... produced the strongest advantage for preseason ADP in more than a decade (since 2014), so the theory isn't exactly off to a great start.

Verdict: still monitoring (but probably not).

Over the biggest sample available, preseason ADP predicts stretch performance almost exactly as well as early-season performance does. Over smaller sub-samples, ADP predicts stretch performance almost exactly as well as early-season performance does. Both perform roughly equally whether you're looking at surprises or disappointments. Both perform roughly equally whether you're looking at early-round players or late.

I've been doing this for fourteen years, and I have yet to find a split where preseason ADP and early-season performance weren't equally good at predicting rest-of-year performance. The first four weeks of the season feel like they're incredibly meaningful, but the truth is that they only tell us as much as we already knew over the offseason.

(Of course, just comparing preseason ADP to early-season results is a false dilemma; as you can see, we're better off basing our expectations on a blend of the two. A strict 50/50 mix of both sources predicts rest-of-year performance substantially better than either source alone. And when I've tested Footballguys' rest-of-year projections using this same methodology, they've performed even better still.)

This, to me, is the ultimate example of regression at work. Whether our players have been disappointments or pleasant surprises to this point of the year, it's important to remind ourselves that in the long run, everyone still trends back toward expectations.

Follow @AdamHarstad Follow @Footballguys