As of October 15th last year, Le’Veon Bell had played six games with the Pittsburgh Steelers and scored just one measly touchdown— a touchdown that came all the way back in week 1. Many owners were beginning to despair, wondering if he would be the next Warrick Dunn or Tiki Barber, two phenomenal backs who historically struggled to reach the end zone.
Undeterred, I mentioned on Footballguys' message board, (which, as an aside, is a fantastic resource well worth checking out), that backs like Le’Veon Bell had historically scored about 7.5 touchdowns over the final ten games of the season.
Sure enough, Bell broke his touchdown drought the very next week. He reached the end zone again in his 9th game, and then scored at least one touchdown in five straight games through November and December. He wound up exceeding even my seemingly optimistic projection, scoring ten touchdowns in his final ten games.
As of November 19th last year, Julio Jones had played phenomenally through ten games with Atlanta, but was barely clinging to the bottom of the top ten fantasy receivers because he only had three touchdowns. More troublingly, he’d gone seven straight weeks without hitting paydirt, and many owners were beginning to despair, wondering if Julio might be another Jimmy Smith, a terrific receiver who rarely reached the end zone.
Undeterred, I mentioned that receivers like Julio Jones had historically scored about 3.4 touchdowns over the final six games of the season. Indeed, Julio broke his scoring drought that very week with a touchdown, then added one more in each of the next two games before getting injured. In total, he finished with three touchdowns in five games, double his rate over the first half of the year.
Am I a wizard? Some mystical future-predicting robot? Should you start calling me "the Predictor" and looking around for opaque boxes? Hardly. I’m just a guy with access to a historical database and a strong belief in the power of regression to the mean.
What did I mean, for instance, when I said “backs like Le’Veon Bell”? In this case, I meant backs who through six games had 600+ yards from scrimmage, (Bell had 793), but two or fewer touchdowns, (Bell, recall, had one). Historically, backs with a lot of yards and very few scores early in the year rebounded dramatically and started scoring at a much brisker pace going forward.
And when I discussed “receivers like Julio Jones”, I was talking about receivers who topped 800 yards through 10 games, (Julio had 912), but averaged fewer than one touchdown for every 300 yards gained, (Julio was at one per 304 yards). Historically, receivers with a lot of yards and very few scores midway through the year rebounded dramatically and started scoring at a much brisker pace going forward.
Predicting Regression for Fun and Profit
One of my favorite mantras is “touchdowns follow yards”. This is what that looks like in practice. In 2014, the top 36 fantasy receivers in standard scoring averaged 140 yards for every touchdown scored. There are sometimes persistent outliers, but they are rarely large, and even outliers tend to regress towards the mean.
Dez Bryant, for instance, is the perfect example of a receiver who seems to have a “knack” for reaching the end zone. In 2014 he scored one touchdown for every 82.5 yards gained, which is phenomenal.
But even outliers operate within limits. Jerry Rice, Randy Moss, Terrell Owens, Cris Carter, and Marvin Harrison are the five most prolific touchdown-scoring receivers in history. All have reached the end zone at least 28 times more than any other wide receiver.
Rice, for his career, averaged a touchdown every 116 yards. Moss averaged one every 98 yards. Owens had one very 104 yards. Carter, (who famously “only caught touchdowns”), had one score every 107 yards. Harrison reached the end zone once for every 114 yards.
It seems the biggest outliers still struggle to move noticeably beyond 1 score for every ~100 yards. And, indeed, Dez Bryant’s career average is one touchdown for every 98 yards, well within range of that apparent lower bound.
On the other end of the spectrum, many great receivers were prolific at gaining yards, but struggled to convert those yards into touchdowns. And, once again, these outliers seem to operate within some sort of upper bound.
The most prolific “yardage-heavy” receivers in history were James Lofton (one TD per every 187 yards), Henry Ellard (one per 212), Andre Johnson (one per 208), Steve Smith (one per 182), and Torry Holt (one per 181). Art Monk, who was something of the anti-Cris Carter, (“all he does is catch 8-yard hitches"), averaged one score per every 187 yards. Jimmy Smith had one for every 183.
Indeed, to find a receiver who, for his career, averaged more yards per touchdowns than Ellard’s 212, you have to go all the way down to Jeff Graham at 82nd on the all-time receiving yardage list. Graham was a journeyman who played for five teams in his eleven year career and never made the pro bowl.
To best Moss’s one touchdown per 98 yards average, you’d have to go all the way down to Don Hutson at 90th on the all-time receiving list. Hutson, you might recall, was an outlier in more ways than just this one. The next name to top Moss in converting yards into touchdowns is Sonny Randle, who played so long ago that his positional designation was “split end”, and who currently ranks 175th on the all-time receiving yardage list.
Bigger outliers do exist at running back, but it’s largely due to specialization and usage differences. For example, on the 2005 Atlanta Falcons, Warrick Dunn averaged 409 yards per touchdown scored, while T.J. Duckett reached the end zone once for every 55 yards gained. The fact that one team featured two such extreme outliers was not some cosmic coincidence, but a simple logical follow: Dunn’s touchdown numbers were so low because he was rarely used in the red zone. Duckett’s yardage totals were so low because he was rarely used outside of the red zone.
Barring a clear split like that, though, running backs can and do regress just as easily as receivers do— as evidenced by Le’Veon Bell and his low-scoring cohort through six games.
So if you want to look like some sort of crazy future-predicting robot, look for players who are converting yards into touchdowns at an abnormally high or an abnormally low rate… and predict they will cease to do so going forward. It’s the easiest bet in fantasy football.
A Brief and Perhaps Interesting List
Here is a list of the top 100 players of 2015 in total fantasy points (standard scoring), sorted by yards per touchdown. A complete list, if you will, from Hill to Hilton. Use it as you see fit.
|86||Melvin Gordon III||5||5||31.6||6.32||71||270||0||13||86||0||Undefined|
|36||Willie Snead IV||6||2||49.6||8.27||0||0||0||26||436||1||436|
|70||Duke Johnson Jr||5||3||35.3||7.06||40||121||0||21||172||1||293|
|6||Mark Ingram II||6||4||75.7||12.62||88||307||4||27||230||0||134.25|
|69||Marvin Jones Jr||5||3||35.5||7.1||2||9||0||15||226||2||117.5|
|58||Ted Ginn Jr||4||4||38.6||9.65||0||0||0||12||206||3||68.7|
One Thought Worth Noting
In my post on Julio Jones, I found that the other receivers in his high-yardage comps averaged 65/950/2.6 over the first ten games and 37.1/506/3.4 over the last six games. While their touchdown rate rebounded dramatically, their yard-per-game average held up fairly well (falling from 95 over the first ten games to 84 over the final six).
I also added a comparison for Brandon Marshall, who had 586 yards and 8 touchdowns through 10 games, a ratio of 73 yards for every touchdown. Players in Marshall’s low-yardage comps averaged 52/741/11.3 over the first ten games and 27/394/3.6 over the final six. Their low yard-per-game average did not rebound in the same way that the high-yardage sample’s touchdown-per-game average did. Indeed, Marshall’s comps saw their ypg decline even further, from 74 down to 66.
The difference between the two samples is especially pronounced if you compare the samples side-by-side. The high-touchdown group scored 8.7 more touchdowns, on average, than the low-touchdown sample over the first ten games. They scored just 0.2 touchdowns more over the final six games. (And this was despite the “high-touchdown” receivers including Randy Moss in 2007 and Mark Clayton in 1984, the year each receiver set the receiving touchdowns record).
Over the first ten games of the season, the “high touchdown” group outscored the “low touchdown” group by 31.3 fantasy points in standard scoring, good for a 3.1 point per game advantage. Over the final six games of the season, the “low touchdown” group outscored the “high touchdown” group by 10 points, good for a 1.7 point per game advantage. That’s a swing of nearly five points per game, a shockingly large turn of events. (In PPR scoring, the swing was even larger, still.)
Trading a “high-touchdown” receiver for a “high yardage” receiver who had scored slightly fewer fantasy points would in many cases have been a season-defining move, as both samples maintained their yardage pace and regressed to the mean on their touchdown pace.
While touchdowns have a tendency to follow yards, yards do not have the same tendency to follow touchdowns. Or, to phrase it for a more social-media-savvy crowd… touchdowns follow yards, but yards don’t follow back.
More articles from Adam HarstadSee all
More articles on: DynastySee all
Waivers of the Future: Week 8 - Simpkins
Cracking FanDuel: Week 7 - Knotts
The New Reality No.152: Quarterback Touchdown Regression Candidates - Parsons
More articles on: StatsSee all
IDP Matchups to Exploit and Avoid: Week 7 - Larkin
For The Win: Week 7 - Larkin
Normalized Strength of Schedule: Week 7 - Lee