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 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.

Most importantly, because predictions mean nothing without accountability, I report on all my results in real time and end each season with a summary. Here's a recap from last year detailing every prediction I made in 2022, along with all results from this column's six-year history (my predictions have gone 36-10, a 78% success rate). And here are similar roundups from 2021, 2020, 2019, 2018, and 2017.

The Scorecard

In Week 2, I broke down what regression to the mean really is, what causes it, how we can benefit from it, and what the guiding philosophy of this column would be. 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 explained that touchdowns follow yards, but yards don't follow touchdowns, and predicted that high-yardage, low-touchdown receivers were going to start scoring a lot more going forward.

In Week 5, we revisited one of my favorite findings. We know that early-season overperformers and early-season underperformers tend to regress, but every year, I test the data and confirm that preseason ADP is still as predictive as early-season results even through four weeks of the season. I sliced the sample in several new ways to see if we could find some split where early-season performance was more predictive than ADP, but I failed in all instances.

In Week 6, I talked about how when we're confronted with an unfamiliar statistic, checking the leaderboard can be a quick and easy way to guess how prone that statistic will be to regression.

In Week 7, I discussed how just because something is an outlier doesn't mean it's destined to regress and predicted that this season's passing yardage per game total would remain significantly below recent levels.

In Week 8, I wrote about why statistics for quarterbacks don't tend to regress as much as statistics for receivers or running backs and why interception rate was the one big exception. I predicted that low-interception teams would start throwing more picks than high-interception teams going forward.

In Week 9, I explained the critical difference between regression to the mean (the tendency for players whose performance had deviated from their underlying average to return to that average) and the gambler's fallacy (the belief that players who deviate in one direction are "due" to deviate in the opposite direction to offset).

In Week 10, I discussed not only finding stats that were likely to regress to their "true mean", but also how we could estimate what that true mean might be.

In Week 11, I explained why larger samples work to regression's benefit and made another yards per carry prediction.

In Week 12, I used a simple model to demonstrate why outlier performances typically require a player to be both lucky and good.

In Week 13, I talked about how a player's mean wasn't a fixed target and predicted that rookie performance would improve later in the season.

In Week 14, I mentioned that hot and cold streaks are mostly a mirage and that all players tend to regress strongly toward their full-season averages.

In Week 15, I looked at the disheartening finding that even the best teams only win a title 30-40% of the time.

In Week 16, I explored the tension between predictions that were interesting and predictions that were likely and how I try to walk the line between both.

In Week 17, I discussed how we can leverage the principles of regression in dynasty leagues by betting on talent and against situation.

Back in Week 7 I predicted that just because the passing production was a massive outlier didn't mean it was destined to regress. Teams did, in fact, finish with the lowest passing yard per game average since 2009, a full 8 yards per game below last year's average (which was already the second-lowest value since 2010). But I predicted not just that values would remain extremely low (compared to recent historical trends), but that they'd actually trend even lower than they were over the first half of the season, and that didn't happen, so this goes down as a loss.

Our last projection went much better for us. In Week 14, I produced the sample of players who were the "hottest" entering the fantasy playoffs. Collectively, they averaged 12.08 points per game over the full season but 16.09 over the last month, a 33% improvement. And then I predicted that the hot streak was a mirage and they group would perform at least twice as close to their full-season average as their recent level. How'd they do since the prediction? 12.07 points per game-- About as close to the target as we could have gotten.

18 out of the 31 players produced at or below their full-season average. 5 more produced slightly above their full-season average, but significantly below their hot streak, still. 6 players managed to mostly maintain their elevated level of play. And 2 players even managed to improve on their hot streak and take their game up another notch entirely. Unsurprisingly (given our penultimate prediction), those last two players were both rookies, the only class of players who consistently perform better late in the season.

Rashee Rice had a full-season average of 11.7 and a "hot streak" average of 14.3 points per game. Over the last month, that rose to 18.4. And Jayden Reed has a full-season average of 11.5 but a "hot streak" average of 14.8. Over the last month, that rose all the way to 21.5. Among receivers, only CeeDee Lamb and Amari Cooper averaged more fantasy points per game than Reed.

Our Final Report Card

To wrap up the season, I wanted to look back not just at this year's predictions, but at every prediction since Regression Alert launched in 2017. Remember, I'm not picking individual players, I'm just identifying unstable statistics and predicting that the best and the worst players in those statistics will regress toward the mean, no matter who those best and worst players might be.

Sometimes this feels a bit scary. Predicting that stars like Christian McCaffrey or CeeDee Lamb, in the middle of league-winning seasons, are going to start falling off is an uncomfortable position. But looking back at our hit rate over time makes it a bit easier to swallow.

Top-line Record

• 2017: 6-2
• 2018: 5-1
• 2019: 7-2
• 2020: 6-1
• 2021: 8-1
• 2022: 4-3
• 2023: 5-3
• Overall: 41-13 (76%)

The One-Off Misses

2017 Passing Yards per Touchdown Part 1
2017 Passing Yards per Touchdown Part 2

In our first prediction, Group A was outscoring Group B by 13%. I picked a bad four-week span to make the prediction, as they outscored Group B by 17% over our prediction span, but over the full season, that fell to just 3%; solid regression, but not enough to count the prediction as a win. When I repeated the prediction later in the season, it once again went poorly. My takeaway from this experience was that quarterback yard-to-touchdown ratios were much more skill-based than running back or receiver ratios (an idea that's backed up by looking at the leaderboard in the statistic), so I've stopped making this prediction anymore.

2018 Yards per Target

Just like with the last miss, I tried to make a prediction out of a statistic that had a large skill element to it. Over the full season, Group A's edge fell from 16% to 7%, which was at least movement in the right direction but not enough to qualify as a win. Once again, I've stopped trying to figure out clever ways to make this prediction work because the skill signal is just too strong, which means the movement going forward tends to be far less dramatic, and the prediction is a bit less reliable. (We did log one hit to offset this one miss before I discontinued the prediction.)

2019 Patrick Mahomes II Touchdown Regression

I knew going into this prediction that it wasn't a great bet; in fact, I preceded the prediction with 18 paragraphs and 4 charts detailing the three biggest issues with the prediction I was about to make, then compounded the issues by breaking best practices again to make a prediction about a single player rather than a large sample (where the ups and downs would have more chance to even out), and broke them a second time by specifically choosing my player rather than sticking with whoever happened to be most extreme in the statistic I was betting on regressing. Then, when the original prediction was lost in part because Mahomes was injured during the sample, I doubled down when he returned from injury and ran it again; this prediction was responsible for both of my losses that year. Really just a disaster from start to finish, with a pair of humbling and well-deserved losses to show for it.

2020 Point Differential vs. Record

I paired teams who had the same record despite wildly different point differentials and predicted that the teams that were winning by bigger margins would win more games going forward than the teams that were winning by smaller margins. Not only did that prediction not work out over the four-week sample, but extending it out through the entire season didn't help any; our Group A teams have actually won one more game than our Group B teams since the prediction. The lesson I take away from this failure is... nothing. Sometimes, predictions fail because I got greedy or made an ill-advised design choice. But sometimes, we just get unlucky. In the future, I'd be happy to make this bet again.

2021 Kickers (Offense vs. Defense)

The point I wanted to make is that a team's offense predicted future performance more than the opposing team's defense. It's a point I've made in the past using offensive and defensive production directly, but this time I wanted to add a twist on it by focusing on "matchups". I think it's a sound point and would be happy to make the prediction again, but my mistake was focusing on kickers; matchups aren't a big deal in fantasy football, but the positions where they're the biggest deal are kicker and fantasy defense (which actually reinforces the underlying point that offense is more predictable than defense). If and when I run this back, I'll pick a different position to focus on.

2022 Schedule Luck

This is a loss with a good lesson behind it. I think the prediction was actually correct, but when I made it, I noticed discrepancies in the data that shouldn't have been there. I even made a note of it at the time, writing, "There are currently 35 teams with a winning record (10-8 or better) and an all-play percentage of 50% or worse. This is our Group A. On the other end, there are 269 teams with a losing record (8-10 or worse) and an all-play percentage greater than 50%. This is our Group B. I don't know why the second sample is so much larger than the first; this may again be a function of the Victory Point screwing with our data."

In theory, luck should be relatively symmetric, and Group A should have been approximately the same size as Group B. The fact that it wasn't was a red flag for me, but it should have been a bigger one. I think luck regressed, but my data source had bad data, and I couldn't catch it as a result. In the future, I'll work harder to make sure Regression Alert adheres more strongly to the principles of GIGO-- "Garbage In, Garbage Out".

2022 Offensive Identity

This was a prediction I'd made once before with great results (a 75% swing in offensive identity!), but this time, for one reason or another, it just didn't hit (a mere 8% swing). I don't think it was a terrible prediction. I think we probably just got unlucky. I might even try it again in the future. But if I do, I'll do a bit more research first to see if this is really something that regresses as much as I think it should or if we just got lucky the first time we tried this.

2023 Passing Yards per Game

This is another example of getting a bit greedy with a prediction. Not content to merely predict the lowest total the league had seen in more than a decade, I decided to predict a 10-yard drop from the previous season (which was already substantially below recent trends), and instead we only wound up with an 8-point drop. Oh well, nothing ventured, nothing gained.

The Frequent Fliers

Here's the outcome of all of my "Yards per Carry" predictions over the years, with the average at the time of the prediction, the average in the four weeks after the prediction, and the total swing.

• Group A had a 60% lead, Group B had a 16% lead, +76% total swing
• Group A had a 25% lead, Group B had a 16% lead, +41% total swing
• Group A had a 24% lead, Group B had a 4% lead, +28% total swing
• Group A had a 9% lead, Group B had a 23% lead, +32% total swing
• Group A had a 20% lead, Group B had a 30% lead, +50% total swing
• Group A had a 22% lead, Group B had a 23% lead, +45% total swing
• Group A had a 3% lead, Group B had a 36% lead, +39% total swing
• Group A had a 10% lead, Group B had a 4% lead, +14% total swing
• Group A had a 24% lead, Group B had a 25% lead, +49% total swing
• Group A had a 42% lead, Group A had a 10% lead, +32% total swing
• Group A had a 10% lead, Group A had a 19% lead, -9% total swing

Sadly, we surrendered our perfect record this year with not one but two losses. The first was just a case of getting a little bit greedy; the swing from Group A to Group B (32%) is actually the median swing for this prediction, but I chose a Group A with such a large lead that it wasn't quite enough to pull Group B ahead. The second was... well, the second was a reminder that even the best bets just go terribly wrong sometimes.

We can't directly compare the total swings since the sample sizes vary so much (a 30% swing over a large sample might be more impressive than a 50% swing over a small one). But we can combine every Group A back at the time of the prediction into one giant "high-ypc" group and combine every Group B back into a massive "low-ypc" group. If we do that, we find that our Group A backs had 3366 carries for 18,802 yards when we made our predictions, a 5.59-yard per carry average. Our Group B backs, by contrast, had 4298 carries for 16,490 yards, or 3.84 yards per carry.

In the four weeks following our predictions, Group A backs rushed 3185 times for 14292 yards, while Group B rushed 4074 times for 18,223 yards. Group B had 27% more carries before the prediction and 28% more carries after the prediction-- their workload advantage was very stable. Group A, by contrast, had a 46% yard per carry advantage before the prediction, but that fell to a 0.4% advantage after the prediction (4.49 to 4.47). Overall, we've gone from Group A rushing for 14% more yards to Group B rushing for 28% more.

Bad year for us or not, yards per carry still isn't a thing.

Here's the outcome of all of my "Yard to Touchdown Ratio" predictions over the years. (Where necessary, I've reworked some predictions to adhere to our traditional "Group A vs. Group B" format. This is a purely cosmetic change for comparison; the underlying data remains untouched.)

• Group A had a 28% lead, Group B had a 1% lead, +29% total swing
• Group A had a 21% lead, Group B had an 8% lead, +29% total swing
• Group A had a 7% lead, Group B had a 20% lead, +27% total swing
• Group A had a 28% lead, Group B had a 23% lead, +51% total swing
• Group A had a 26% lead, Group B had a 4% lead, +30% total swing
• Group A had a 23% lead, Group B had a 47% lead, +70% total swing
• Group A had a 22% lead, Gorup B had a 23% lead, +45% total swing
• Group A had a 2% lead, Group B had a 40% lead, +42% total swing
• Group A had a 15% lead, Group B had an 11% lead, +26% total swing
• Group A had a 9% lead, Group B had a 13% lead, +22% total swing
• Group A had a 10% lead, Group B had a 19% lead, +29% total swing
• Group A had a 3% lead, Group A had a 12% lead, -9% total swing
• Group A had a 7% lead, Group B had a 38% lead, +45% total swing

Our perfect streak finally ended in 2022, but an early win this year got us back on the right track. Predictions that touchdowns will follow yards are 12-1, with a median swing of 29%. 12 out of 13 predictions have produced a swing of at least 22%.

I've said for years that yard-to-touchdown ratio is my favorite statistic precisely because it's such a reliable regression target but no one else pays any attention to it. It's also the reason we're here today; this entire column was inspired by a pair of articles I wrote on this ratio back in 2015. Everyone knows sometimes players score "too many" or "too few" touchdowns, but virtually none of the discussion links that judgment to their yardage profile. As you can see, that's exactly the link we should be making-- or as I put it originally: touchdowns follow yards (but yards don't follow back).

Here are the various other miscellaneous (successful) predictions from the past four seasons

• Group A had 16% more yards per target, Group B had 11% more yards per target, +27% total swing
• Group A had 17% fewer interceptions, Group B had 57% fewer interceptions, +74% total swing
• Group A had 13% fewer interceptions, Group B had 17% fewer interceptions, +30% total swing
• Group A had 25% fewer interceptions, Group B had 11% fewer interceptions, 36% total swing
• Group A had 20% more kicker points per game, Group B had 36% more kicker points per game, +56% total swing
• Group A had 42% more rushing TDs per game, Group A had 33% more passing TDs per game, +75% total swing
• Group A recorded 33% more interceptions, Group B recorded 24% more interceptions, +57% total swing
• Group A recorded 65% more interceptions, Group B recorded 50% more interceptions, +115% total swing
• Group A won 4% more fantasy matchups, Group B won 19% more fantasy matchups, +23% total swing
• Group A averaged 93% more yards per route run, Group B averaged 49% more yards per route run, +142% total swing

And general regression predictions that didn't follow the typical "Group A vs. Group B" format, instead predicting unidirectional regression for a single group.

• "Extreme" offenses and defenses regressed 11% toward the league average performance, as predicted.
• Defenses regressed 12% more than offenses, as predicted.
• Group A averaged 14% more passing yards per game, Group A continued to average 28% more passing yards per game, as predicted.
• Margin of Victory settled between 9.0 and 10.5 points per game, as predicted.
• Rookie receivers increased their production over the final month of the season, as predicted.
• That increase was broad-based rather than driven by outliers, with 64% seeing an improvement, as predicted.
• "Hot" players regressed 108% of the way back to their full-season averages, as predicted.
• "Hot" players regressed 75% of the way back to their full-season averages, as predicted.
• "Hot" players regressed 133% of the way back to their full-season averages, as predicted.
• "Hot" players regressed 100% of the way back to their full-season averages, as predicted.

I appreciate you all reading along this season. The whole point of this column is to convince you that regression to the mean is real, it's implacable, and it's actionable with very little effort on our own part.

Accountability is crucial to making that point, which is why I go to such great lengths to track and report my results. You don't have to take my word on the subject; you can go back and check my track record for yourself. You can see why I'm such a big believer in the power of regression, and hopefully, you will become something of a believer yourself.