NFL

# Fantasy Football: Regression Candidates Through Week 5

Regression is fantasy football soothsaying when practiced correctly, but that's the key -- it must be practiced correctly. If a player is expected to score one touchdown but instead scores two, that doesn't mean they're expected to score zero touchdowns the following week. It means they're expected to score, as you might have guessed, one touchdown.

In that sense, regression doesn't necessarily mean a player is bad or even overvalued. It means they've produced differently than they're expected to in the future.

Reversing the earlier example, a player who scores zero touchdowns when expected to score one touchdown should expect regression but in a positive sense. Some hate the term "positive regression," but regression is short for regression to the mean -- meaning it can be either positive or negative, depending on which direction the mean is. It can be a mean process. Sorry, but I'm contractually bound to use all three forms of the word mean in this article.

Now on to this week's regression candidates.

### Negative Regression Candidates

Josh Allen, QB, Buffalo Bills

There's often some redundancy in regression, so instead of focusing on that aspect heavily, I'm going to briefly explain what I think is behind Josh Allen's gaudy numbers.

The short answer is play-action. The longer answer is more nuanced, but also mostly play-action. As a refresher: play-action passes are A) more efficient in general and B) less stable to predict than standard dropback passes. For simplicity's sake, I'll also be talking about RPOs (run-pass options) in this context. In general, they're about equally efficient to play-action passes. Prior to Tuesday's game, Allen had a league-high 60% of his passing attempts come off play-action or RPOs, the league's next highest full-time passer was his opponent, Ryan Tannehill at 45 percent. Allen's yards per attempt (YPA) on play-action and RPOs was 9.9, while his YPA on standard dropbacks was 7.6. To be fair, 7.6 is an above average number, it's miles better than what I ever expected out of Allen, but it's not elite like his other efficiency numbers suggested.

I don't think this is necessarily a huge red flag either. Brian Daboll is a great play-caller, and will likely continue running play-action. There are two dangers, however, and we say both on Tuesday night. First, play-action efficiency is less sticky than standard dropback efficiency. Also, trailing teams won't necessarily run a ton of play-action.

Allen will be a good fantasy quarterback with or without passing efficiency, just don't buy tickets for his MVP ceremony yet.

Todd Gurley, RB, Atlanta Falcons

Our quarterback selection ran long, and there's no glaring running back regression candidate, so I'll keep this one short. Todd Gurley is scoring a few too many touchdowns. Gurley saw a slight uptick in targets last week, which is worth monitoring. Otherwise, he's a two-down back seeing just more than 50% of snaps and is on an 0-5, pass-first team. Not ideal.

Mike Evans, WR, Tampa Bay Buccaneers

The slew of two-yard touchdowns has obscured some worrying signs for Mike Evans. He's not even on pace for 1,000 yards this season, and that's in part due to his lack of efficiency. The injuries have potentially slowed him as he's at just 7.7 yards per target, but probably more due to his (average depth of target) aDoT which has dipped from 15.6 in 2018 and 15.1 in 2019 to 9.0 in 2020. Evans is a contested-catch, deep threat receiver. He simply can't provide the same level of efficiency if he isn't being targeted downfield.

Also troubling, Evans has just an 18.3% target share on the season, and that won't be alleviated with the eventual return of Chris Godwin. The touchdown rate will eventually regress, and Evans may be in trouble when that happens.

Robert Tonyan, TE, Green Bay Packers

Robert Tonyan had the week off, so it's probably worth running it back this week and reminding folks to keep expectations in check and not go chasing touchdowns. Tonyan is still an athletic tight end with an elite quarterback, and thus very worth having. Just don't expect that he'll instantly be Travis Kelce. Don't start him over Zach Ertz or something like that.

### Positive Regression Candidates

Lamar Jackson, QB, Baltimore Ravens

I wrote about Lamar Jackson's weird play-action splits last year. I'm proud of this article, but I regret my choice in title. If I could rename it today it would be something like "Studying Play-Action Splits Through Lamar Jackson." We don't have a ton of data on play-action passing yet, but in this case study, we find that Jackson went from his play-action and RPO passes being roughly half as efficient as standard dropbacks in 2019 to be where we'd expect in 2020, meaning he's more efficient on RPOs and play-action than standard dropbacks this year.

We no seem to know, that Jackson isn't uniquely bad a play-action passing, but we also know that standard dropbacks are more predictive and he's dropped from 10.2 YPA on standard dropbacks in 2019 to 6.6 YPA in 2020. We should expect this number to improve if not reach 2019's heights. Since yards drive most fantasy production this will obviously be a huge boost.

Joe Mixon, RB, Cincinnati Bengals

Joe Mixon may have scored all his touchdowns in one game, but don't let that fool you he's still underperformed touchdown expectation. Expect touchdowns on a more consistent basis, with the occasional big game mixed in for Mixon, especially with his newfound receiving game involvement.

Robby Anderson, WR, Carolina Panthers

This is a pro-D.J. Moore column first and foremost, but at some point, it's tough to dispute Robby Anderson's role in the offense, and with that alpha status and associated yardage we should expect more touchdowns from a certain Sir Purr skeptic.