Fantasy analysis says “touchdown regression” so often that it can become a substitute for analysis. We wanted a falsifiable version: does an opportunity-based expected-touchdown baseline predict next season better than a player’s raw touchdown total?
Across 551 consecutive WR/TE season pairs, it did. Carrying forward actual touchdowns produced 2.61 touchdowns of mean absolute error. Carrying forward expected touchdowns cut that to 2.38 — an 8.9% reduction.
The result in one picture
Players in the highest overperformance quintile scored 8.2 touchdowns in year one and 5.4 the next: a decline of 2.8 touchdowns. The biggest underperformers moved the other way, from 3.0 to 4.3.
That is the useful core of regression. It is not “high-touchdown players are bad.” It is “the part of touchdown production unexplained by repeatable opportunity should receive less weight.”
What we modeled
We aggregated nflverse regular-season receiving data for wide receivers and tight ends with at least 40 targets from 2014–2024. For every test season from 2018 onward, the model was trained only on earlier seasons. Five same-season opportunity signals — targets, receptions, yards, air yards, and receiving first downs — estimated a touchdown baseline. Ridge regularization reduced instability from correlated volume measures.
Population WR/TE, ≥40 targets
Training Expanding window
Tests 2018–2024
Seed 4,444
The next-season test compared two deliberately simple forecasts against the same outcome:
- next season equals this season’s actual touchdowns; and
- next season equals this season’s expected touchdowns.
The paired MAE improvement was 0.23 touchdowns, with a 95% paired-bootstrap interval of 0.09 to 0.38. Those 5,000 resamples quantify uncertainty inside this one evaluation sample; they are not independent replications. Both forecasts still missed by more than two touchdowns on average, so this is a modest benchmark improvement rather than a complete projection model.
Every quintile, not just the headline
| TD − xTD group | N | Current TD | xTD | Next TD |
|---|---|---|---|---|
| Most under | 111 | 3.0 | 5.6 | 4.3 |
| Under | 110 | 3.5 | 4.7 | 4.4 |
| Near expected | 110 | 4.3 | 4.4 | 4.4 |
| Over | 110 | 5.4 | 4.6 | 4.7 |
| Most over | 110 | 8.2 | 5.0 | 5.4 |
How to use it
At the population level, large touchdown-over-expected gaps can support a weak tiebreaker—not an individual projection adjustment with a known magnitude. If two players have similar projected opportunity, the evidence supports giving less weight to the one whose prior production depended more on outlier conversion. Do not mechanically fade elite players with elite volume, and do not turn a historical baseline into a current projection without team, quarterback, health, and betting-market context.
Regress the conversion. Preserve the opportunity. Then update both for the player’s current environment.
Where this can break
The test includes only players who returned to at least 40 targets, so it describes volume survivors. Of 950 walk-forward predictions, 551 became evaluable pairs, 257 fell below the next-season target threshold, and 142 await a completed outcome season. Selection can affect both tails differently, and the net direction of bias in the forecast comparison is uncertain. Season totals also hide red-zone usage and route context. The model is a benchmark, not a causal estimate of touchdown “skill.” Finally, the public source snapshot ends with 2024; this is not a current player-ranking model.
Reproducibility and provenance
Sources: nflverse player stats (CC BY 4.0 (nflverse compilation; underlying NFL data retain owner terms)). The run manifest records each source hash, code version, random seed, model coefficients, walk-forward predictions, bootstrap outputs, five reviewer findings files, and the publication decision.