In Week 2 of last year, Julian Edelman out-scored Tyler Higbee, with 28.4 fantasy points to 25.9. That’s despite the fact that while Edelman caught 8 of 11 targets for 179 yards, Higbee managed only 54 yards on 5 catches. The difference? Touchdowns — Higbee scored three times, Edelman scored zero.
Clearly, touchdowns – worth 60 yards in standard leagues and/or six catches in PPR leagues – are extremely valuable in fantasy football, but touchdowns are also rare events, mostly random, and extremely hard to predict. At each position, touchdowns were far less sticky (showing a much lower correlation year-over-year) than targets, receptions, carries, or yards. We also know that players who ranked highly in touchdowns in one year tend to regress closer to the mean in the following season.
Over the past decade, there are 143 instances of a flex-eligible player recording 10 or more touchdowns in a single season and playing in the following season. Of those instances, only 11% saw an increase in touchdowns the following year, and, on average, each player lost about 5.3 touchdowns from their prior-season total.
So, we know touchdowns are extremely important for fantasy football but extremely difficult to predict. And because so many people just draft off of last year’s results, we know it’s valuable to pinpoint looming touchdown regression candidates for fantasy. But what’s the best method for evaluating this?
A player’s talent does play a small role in scoring touchdowns, but for the most part, touchdowns are merely a function of a player’s opportunity. The more opportunities you receive and the more opportunities you receive closer to the end zone, the more likely you are to score a high number of touchdowns. And red-zone metrics, though frequently cited, aren’t a very good measure of near-end-zone opportunities, nor are they very good at predicting touchdowns. Why? Because, with red-zone carries, you’re going to be grouping carries from the 1-yard line (53.8% chance of scoring) in with carries from the 17-yard line (3.6% chance of scoring) and treating them the same. A much better stat would weight each carry (and target) at each yard line by the actual degree of scoring probability.
Mike Clay first came up with this idea in 2013, and created a metric called oTD to improve upon red-zone metrics as a predictor for touchdowns. With Clay’s oTD as the inspiration, I attempted to calculate expected touchdowns (XTDs) using a similar methodology to Clay’s. Basically, I’m looking at each carry by distance from the end zone (and weighting each carry according to the average scoring rate on all identical carries over the past decade) and looking at each target by distance from the end zone and depth of target (and weighting each specific target based on the average scoring rate on all identical targets over the past decade).
The sum of these numbers gives us a player’s total expected touchdowns, or how many touchdowns we should expect a perfectly average player to score on an identical workload. We can contrast this with a player’s actual touchdown total to measure how lucky or unlucky they were in a given season. And, again, know that I say “unlucky” as opposed to “inefficient” for a reason, and that’s because…
After testing this, I found this stat – expected touchdowns (XTD) – to be even stronger at predicting touchdowns in the following season as raw touchdowns (0.283 to 0.275 RSQ) and far stickier year-over-year (0.382 to 0.276 RSQ).
Of the top-25 seasons by (positive) touchdown differential this past decade, only one player scored more touchdowns in the following season, and on average players saw their touchdown total decline by 52% in the following season. Of the bottom-25 seasons (worst touchdown differential), 18 of 25 improved in the following season, and, on average, by 63%. As you can see, clearly, at the polar extremes, a massive touchdown regression is extremely likely (hello Derrick Henry, Aaron Jones, Leonard Fournette).
Without further ado, by contrasting expected touchdowns with actual touchdowns, here are the players most likely to regress in the touchdown department next season: Notes: The introduction to this article was written by Scott Barrett. Everything subsequent to this note was written by Jake Tribbey. You can find XTD data here in our Expected Fantasy Points tool.
Top Regression Candidates 
Top (Positive) Regression Candidates
The Outliers
As noted in the introduction, raw volume plays a much more significant role in touchdown production than player skill. That’s true, but there are certain players who can consistently out-score their touchdown-based expectation year over year. Such players are rare, but they do exist. Tyreek Hill, Alvin Kamara, and A.J. Brown are among those players, the rare hyper-efficient outliers I think we can expect to continue to outproduce their expectation. Since entering the league in 2017, Kamara has scored 21.0 touchdowns more than his expectation implies, easily leading the NFL over that span. Since entering the league in 2016, Hill has scored 25.2 more touchdowns than his expectation implies. That’s also a league-high, with no other player over this span coming anywhere near Kamara or Hill. Similarly, A.J. Brown ranks behind only Derrick Henry over the past two seasons (+11.1). He outscored his expectation by +5.0 touchdowns in 2019 and then +6.1 in 2020. Both numbers ranked top-4 each season, and likely confirm that Brown is a “special,” possibly “generational,” talent at the WR position, rather than a player who is just getting tremendously lucky with his opportunities. Still, with a smaller sample at our disposal, it’s less of a “sure thing” than with Hill or Kamara.
Will these players regress? Sure, maybe a bit, but I think they’ve proven that the mean they’re regressing to is quite a bit higher than that of a league-average player.
Davante Adams, WR, Green Bay Packers
XTD: 11.6, TD: 18, Diff: +6.4
Davante Adams averaged 26.3 FPG during the fantasy season last year, which was the most by any WR since 1960. He scored the 3rd-most touchdowns by any WR in any season all-time, and did so despite missing 2.5 games. Even if Adams didn’t score a single TD last year, he still would have finished as the WR5 by FPG. He led all WRs in YPRR (2.96), PFF grade (92.2), and passer rating when targeted (136.9). Obviously, after such a historically great season, a regression to the mean is inevitable, even if he is a near-outlier like Kamara and Hill.
Adams may be talented enough to put up another season with similar TD output to 2020, but certainly not without reigning MVP Aaron Rodgers. He fed Adams 34% of the team’s targets (44% in the red zone) and more importantly did so with outstanding accuracy. Per PFF’s QB Annual, Rodgers was top-7 in accurate pass % and 2nd in catchable pass rate last season. With Adams as the main beneficiary, he stands to lose the most from a production standpoint if Rodgers isn’t in Green Bay come Week 1.
Robert Tonyan, TE, Green Bay Packers
XTD: 5.1, TD: 11, Diff: +5.9
Robert Tonyan is a screaming touchdown-regression candidate, probably the biggest one at the position since 2018 Eric Ebron or 2015 Tyler Eifert. Tonyan finished the 2020 season as the overall TE3 despite ranking just 24th in targets (59). Had he scored only his expected allotment of touchdowns, he would have dropped from TE5 in FPG (11.0) to TE18 (8.8). Remarkably, Tonyan had more touchdowns (11) than he had incompletions thrown his way (7). Since 2008, there are only five occasions of a TE exceeding his expected TD output by a greater margin than Tonyan did last season. He averaged 2.99 fantasy points per target, which was the most by any 50-target TE in any season all-time. On one hand, an efficiency regression is beyond inevitable. These numbers are impossibly absurd, even with Aaron Rodgers’ MVP-play factored in. On the other hand, his 2020 season was so impressive, so efficient, that an increase in volume seems inevitable.
Robert Woods, WR, Los Angeles Rams
XTD: 3.3, TD: 8, Diff: +4.7
Robert Woods outscored Cooper Kupp in FPG 15.4 to 14.1 last season, but had they been equally lucky in the touchdown department, Kupp would have had the better year, outscoring Woods 14.4 to 13.7. While the XTD data looks favorably upon Kupp, I’m not so sure the QB situation does.
Jared Goff has prioritized targeting the slot like no other QB over the last four seasons, and with Matthew Stafford now taking over under center, I’d expect a modest drop-off in slot targets, which serves to benefit Woods more so than Kupp. I discuss this in more depth here, but Woods has lived in the intermediate 10-19 yard range, and that’s an area Stafford has consistently attacked throughout his career, always ranking top-15 in intermediate targets. I don’t expect Woods to catch 100% (4 of 4) of his end zone targets like last season, so he may take a step back in the TD department, but I’m still banking on him outperforming Kupp (in FPG) with Stafford at QB.
Adam Thielen, WR, Minnesota Vikings
XTD: 9.5, TD: 14, Diff: +4.5
Adam Thielen had a serious outlier of a season when it came to both end zone luck and volume, as he scored 13 of his TDs on 20 end zone targets. Thielen’s 1.33 end zone targets per game ranked 2nd in the NFL, but it was also a drastic increase on the 0.8 end zone targets per game he saw in 2019, or the 0.63 per game he earned in 2018. Thielen’s 65% end zone conversion rate put the league average of 42% to shame, and was also drastically higher than his career mark of 41%. He saw 50% of the team’s total end zone targets and, despite nearly identical red zone workloads, out-targeted his uber-talented teammate, Justin Jefferson, 20-8 in the end zone.
Is it possible Thielen developed into one of the league’s best end zone receiving threats in his age 30 season? Sure, I guess. But what’s much more likely is that he was over-targeted in the end zone because of his familiarity with QB Kirk Cousins and he got lucky by converting an unusually high number (for him or nearly any other WR) of those targets into TDs. For 2021, I’d expect him to land much closer to the 13.8 XFP per game he averaged last year than the 16.8 FPG he ended up with.
Antonio Gibson, RB, Washington Football Team
XTD: 6.6, TD: 11, Diff: +4.4
Antonio Gibson has been one of the offseason’s most buzzed about players, and for good reason. Last year, he delivered the RB17 season by FPG (14.7), but would have fallen to RB26 (12.8 FPG) had he scored TDs as expected. The main reason for this discrepancy was an unsustainably high goal line conversion rate, as Gibson’s 54% TD rate inside the five-yard-line was much higher than the league average rate of 39%.
So, while that’s a number that’s bound to regress in 2021, that should be more than offset by an increase in goal-line volume. Gibson split goal-line work 60/40 with Peyton Barber last year, but was far more efficient than his teammate. Ideally, that moves closer to 100% in Gibson’s favor, and Barber and his laughably inept 2.7 YPC average barely see the field. And, given the massive leap we typically see from 2nd-year RBs and that HC Ron Rivera specifically expects this leap from Gibson, I am anticipating a massive uptick in offensive involvement for the sophomore rusher. That uptick in volume (at the goal line and as a receiver) should more than offset Gibson’s looming TD regression.
J.K. Dobbins, RB, Baltimore Ravens
XTD: 5.1, TD: 9, Diff: +3.9
J.K. Dobbins led the NFL with 6.01 YPC — a mark that tied C.J. Spiller’s 2012 campaign for the 4th-best YPC season by an RB since 2000 (min. 100 rushes). And that wasn’t the only area Dobbins was hyper-successful, as he converted eight of his nine rushing attempts inside the 5 yard-line into TDs, fairing far better than teammate Gus Edwards, who turned his 10 attempts into just three scores. With a near 50/50 goal line split between the backfield pair likely again in 2021, a regression to the mean for Dobbins’ TD output is a near certainty. With that being said, I’d still bet on Dobbins to exceed his rushing expectation consistently going forward, thanks to sharing the backfield with Lamar Jackson.
Primed for Improvement
Ezekiel Elliott, RB, Dallas Cowboys
XTD: 13.4, TD: 8, Diff: -5.4
Since entering the league, Ezekiel Elliott has finished third (21.7), third (20.3), sixth (21.9), fifth (19.5), and last year 16th (14.9) in FPG. 2020 was easily the worst season of his career, posting career-lows in PFF grade (65.3), yards per carry (4.0), yards after contact per attempt (2.8), missed tackles forced per touch (0.18), yards per route run (0.87), and most importantly for this article: touchdowns (8).
So, Elliott no doubt played poorly in 2020, but he was also incredibly unlucky. For one thing, the team lost starting QB Dak Prescott in Week 5. That meant Dallas’ season was effectively over by Week 7, and Elliott’s inefficiency was somewhat excusable due to repeatedly brutal gamescript and opposing defenses stacking the box and then laughing at QBs Andy Dalton (5.34 ANY/A), Garrett Gilbert (4.88), and Ben DiNucci (2.92). For another thing, Elliott was also incredibly unlucky in the touchdown department.
Among the 26 RBs with 10+ carries inside the five yard line, Zeke’s 19% TD rate on those carries ranked 26th, and was well below the league-average rate of 39% or Elliott’s career rate of 43%. His -5.4 expected TD differential was the 6th-worst by any RB since 2008. Had he scored TDs as expected, he would have gone from RB14 in FPG (15.5) to RB7 in FPG (17.6). Unless he’s totally dust (and I don’t think he is), the chances Zeke repeats as the league’s least-efficient goal line rusher are almost zero, and with his health back, his starting QB back, and the Cowboys paying him $13.7 million, Zeke should have no problem returning to mid-range to high-end RB1-status in 2021.
Benny Snell, RB, Pittsburgh Steelers
XTD: 9.3, TD: 4, Diff: -5.3
With 31% of his 111 carries coming in either short yardage or goal line situations, Benny Snell had basically one job: earn tough yards. He didn’t exactly excel in that role, recording the 3rd-worst YPC of any RB (3.32) and struggling at the goal line, converting 15 carries inside the five into just four TDs. Snell was really bad, but there is some good news, at least as it relates to 1st-round pick and new starting RB Najee Harris.
Harris was a TD magnet in college, as his 30 touchdowns in 2020 were the most by any SEC RB ever. He’s an outstanding talent and it’s glaringly obvious the Steelers want to make him their bell cow. At 6’2”, 232, I’d be shocked if that didn’t include a borderline-monopoly on goal line touches, similar to the incredible 98% of red zone carries (100% inside the five) Le’Veon Bell saw as Pittsburgh’s bell cow in 2017. Simply put: Najee Harris is an RB1 in 2021 and must be treated as such.
A.J. Green, WR, Arizona Cardinals
XTD: 6.2, TD: 2, Diff: -4.2
A.J. Green is almost certainly dust. His 66.3 PFF grade was the worst of his career by 7.7 points, and the worst by 15.0 points if you exclude his rookie year. Among the 84 WRs with at least 50 targets in 2020, Green ranked dead last in passer rating when targeted (56.7), yards after the catch per reception (1.8), % of targets caught (46.5%), yards per target (5.0), and EPA per target (-0.18). According to Next Gen Stats, Green generated the lowest average separation (1.7 yards) of any WR last year. He was the 6th-worst WR in on-target catch % (82.1) and YPRR (1.02).
He was horrendous.
A position like WR, which seriously rewards quickness, change of direction, agility, and speed, tends to suffer harsher production drop-offs later in their careers than, say, offensive line, because 30+ year-olds (Green will be 33 when the season starts) always lose their athleticism and never get it back. Green can’t cut or run routes like he used to, and that shows up in his inability to generate separation as he could a few years ago. HC Kliff Kingsbury may be currently calling Green a starter, but I’m not touching him in any format (unless he’s literally free) and I’d be surprised if he was even on Arizona’s roster by the end of the season. Is he due for a positive touchdown regression heading 2020? Maybe technically, but still no, probably not. I think he was more “bad” than “unlucky.”
Devin Singletary, RB, Buffalo Bills
XTD: 5.0, TD: 2, Diff: -3.0
Of the 89 RBs to see 5+ red zone carries, Devin Singletary’s 4.5% TD rate on his 22 carries ranked 81st last year. Singletary saw just seven fewer red zone carries than backfield teammate Zack Moss, but concerningly, Moss out-carried Singletary inside the five 11-6. That follows a pattern, as it appears Bills’ coaches just aren’t comfortable with utilizing Singletary near the end zone, as he was also out-carried by Mr. 3.5 YPC (Frank Gore) 11-2 inside the five in his rookie season. The additional presence of Josh Allen caps the true goal line TD upside of either RB, so it’s hard to view Singletary as having anything close to an appealing ceiling when TDs will be few and far between. Once again, though technically true Singletary might be due for a touchdown regression, I’d love to bet the under on 5.0 for 2021.
Michael Thomas, WR, New Orleans Saints
XTD: 2.6, TD: 0, Diff: -2.6
Since 2016, there are only 20 instances of a WR seeing 40+ targets and failing to catch a single touchdown pass. Last season, Michael Thomas became one of those 20 WRs. He scored 9 in three of his other four seasons, and, well, there’s just no way he scores anywhere near zero touchdowns again in 2021.
Held up by injuries and punching teammates in practice, basically everything that could have gone wrong for Thomas in 2020, did. While many are viewing the potential (only 35% per Vegas) of Taysom Hill starting at QB as a fantasy death sentence for Thomas this season, I’m not. Thomas averaged 16.1 FPG and 9.25 targets per game (32% TMS) when Hill was under center, numbers that both would have ranked top-12 among WRs if sustained over the entirety of the season.
Even historically, the Saint’s have shown that in any and every circumstance, they want to get Thomas the ball. When Teddy Bridgewater started five games in 2019, Thomas posted 23.8 FPG, 10.25 targets, and 105.0 receiving yards per game. We are looking at a fantasy WR1 who’s agnostic to QB play.
Some people may bring up Thomas’s injury last season as a reason not to draft him this year, but outside of suppressing his ADP (making him arguably the best injury discount of 2021), it’s nothing to worry about.
Ready for the best part? Thomas is currently priced at his stone-floor. At WR9 on Underdog and BestBall10s, he’s is an absolute steal in drafts after having the 7th-best WR fantasy season of all-time in 2019.
Jalen Reagor, WR, Philadelphia Eagles
XTD: 3.6, TD: 1, Diff: -2.6
Jalen Reagor primarily suffered from poor efficiency, but it, at least partially, wasn’t his fault. Reagor himself was far from perfect and wound up finishing in the bottom half of WRs in PFF grade (64.0) and yards per route run (1.30), but still managed to tie with Travis Fulgham for the team lead in WR XFP per game (9.1). So how did Reagor only end up scoring 7.2 FPG in 2020? I’m primarily blaming poor QB play.
Why? If you look at all QBs with over 100 dropbacks last year (44 qualifiers), both Carson Wentz and Jalen Hurts ranked bottom-10 in PFF grade, passer rating, and adjusted completion %. If you combined Hurts and Wentz into one player, their 69% catchable target rate would’ve ranked ahead of only Jake Luton.
Reportedly moving to the slot in 2021, I’m expecting Reagor to be helped out by both a higher volume of targets and more passes on target thanks to a lower aDOT. Hurts’ increased comfort at the pro level should help add additional consistency to one of the league’s most inconsistent passing attacks last year. Those factors should lead to more normalized TD scoring for the sophomore WR in 2020.
Robby Anderson, WR, Carolina Panthers
XTD: 5.4, TD: 3, Diff: -2.4
People seem to be forgetting that Robby Anderson was the Panther’s WR1 last year, leading the team with 134 targets. While he was just barely outscored by D.J. Moore in FPG, had Anderson scored TDs as expected he would have led the Carolina WRs in FPG, and jumped from the overall WR29 to WR22. Anderson himself likely fell victim to the Panthers’ sub-par offensive efficiency, as they converted red zone trips into scores at the 5th-lowest rate (51%) and were 28th in passing TDs despite being 17th in passing yards. Both Anderson and this offense as a whole are likely due for an uptick in TD production. At WR31 on Underdog and WR35 on BestBall10s, Anderson is priced at his floor and is one of my favorite WR values.
Clyde Edwards-Helaire, RB, Kansas City Chiefs
XTD: 7.3, TD: 5, Diff: -2.3
The definition of “brutal touchdown luck” early in the season (Weeks 1-6), Clyde Edwards-Helaire found the end zone only once on his 12 rushes inside the 10-yard line, even though he was expected to score 5.3 TDs on those attempts. Had he scored as expected, he would have gone from RB17 (15.9 FPG) over that time period to RB7 (20.2). After Week 6, Edwards-Helaire saw 20+ touches only once, but it’s important to note that Le’Veon Bell joined the roster just in time for Week 7, explaining the volume dip for CEH. With Bell now out of the picture and Andy Reid having historically used his RB1 in a bell cow-like fashion, I’m expecting CEH to see volume comparable to what he saw from Weeks 1-6 (21.3 touches per game) but with more normalized TD luck. That’s a recipe for RB1 production … with top-5 potential.
Myles Gaskin, RB, Miami Dolphins
XTD: 7.3, TD: 5, Diff: -2.3
The 18.3 touches per game that Myles Gaskin averaged throughout the 2020 season ranked 9th among RBs, ahead of notables such as Alvin Kamara, Nick Chubb, and Jonathan Taylor. Gaskin was also 6th in XFP per game (17.6) and 9th in RB snaps per game (45.3). He may not have been recognized as such, but Gaskin was essentially a bell cow in the 10 games he played. It was an impressive season for the 7th Round rookie, but it could have been a league-winning year if not for injuries and poor touchdown luck. Gaskin finished 10th in FPG last year (16.8), but if we neutralized touchdown-luck for all RBs (every RB posted a XTD differential of 0.0), he would have finished 6th (18.0), and just 0.3 behind Aaron Jones (RB5).
With his main RB competition being Salvon Ahmed, Malcolm Brown, and Gerrid Doaks, Gaskin’s a clear favorite to resume that bell cow-esque workload again this season, and with an ADP of RB25, he’s an absolute steal in drafts right now. The Dolphins offense should take a significant step forward given Tua Tagovailoa’s added comfort with the playbook and the WR help from Will Fuller and Jaylen Waddle, and that should mean more TDs and red zone opportunities for the offense, specifically for Gaskin. Personally, I’d be shocked if Gaskin didn’t see an ADP boost in the next few weeks/months as people realize what a drastic market oversight his draft position is.
Cam Akers, RB, Los Angeles Rams
XTD: 6.2, TD: 4, Diff: -2.2
Another victim of poor luck near the goal line, Cam Akers converted just one of nine of his inside the five attempts into scores, fairing far worse than teammates Malcolm Brown and Darrell Henderson who combined to go eight of 20 inside the five. I’m confident we’ll see more normalized goal line luck from Akers this year, but that’s far from the best reason to be excited about the sophomore rusher. In his final six games, he averaged 23.8 touches and 117.0 YFS. Both numbers would’ve been top-5 last year had they been sustained over the entire season.
For 2021, “Cakers” could certainly approach the 22.7 touches and 135.3 YFS that Todd Gurley earned in 2017 and 2018, given how last season ended and that HC Sean McVay has spoken highly of Cakers as a potential bell-cow. Recreating that Todd Gurley role would mean fantasy magic for Akers, as Gurley’s 54 TDs from 2017-2019 were the most by any RB, beating out 2nd-place Christian McCaffrey by 15 TDs. Why? Gurley’s red zone usage was insane. He led the league in red zone carries twice (63 in 2017 and 64 in 2018) and then finished 2nd in 2019 with 56. Over Sean McVay’s four seasons as head coach, the Rams rank 2nd in rushing TDs and 4th in rushing attempts inside the red zone. They’ve been top-5 in red zone rush % every season since 2018.
If Akers can take over 70%+ of the red zone work for LA (which I think he will), he’s going to score TDs, and probably a lot of them. McVay wants a bell cow, and wants to run the ball at a borderline league-leading rate in the red zone. This whole situation shapes up as a dream fantasy scenario for Akers, and could propel him to become the overall RB1 this season. He’s one of my favorite players to draft in every format.
