Slate Stats › NFL › Studies

What actually predicts fantasy production?

Verdict — Opportunity, not efficiency

Playing time. A player’s snap share repeats at 0.98 from one half of a season to the other, and his carries at 0.98. His touchdowns repeat at 0.72 — barely better than a coin flip.

Before asking whether a number is useful, it is worth asking whether it is even a property of the player. A statistic that does not repeat within a single season cannot tell you anything about the next week, however large it looks.

00.20.40.60.810.978Snap share0.982Carries pergame0.962Target share0.943Receptionsper game0.929Receivingyards pergame0.721Touchdownsper gameFull-season reliabilityrepeats at
2,843 player-seasons, 2018–2025. Each bar is how strongly the first half of a season predicts the second for the same player. Everything on the left is how much work he gets. The thing on the right is what he did with it.
StatisticRepeats atReads as
Snap share0.98a property of the player
Carries per game0.98a property of the player
Target share0.96a property of the player
Receptions per game0.94a property of the player
Receiving yards per game0.93a property of the player
Touchdowns per game0.72part signal, part luck

The split is clean, and it is the same split every time. Opportunity — snaps, carries, targets — is close to a fixed property of a player within a season. Efficiency and scoring are not. That is not a claim about talent; it is a claim about sample size. A receiver runs 30 routes a week and scores roughly once a month.

The practical form: project the work first and let the points follow. A player whose snap share just jumped is telling you something about next week. A player who just scored twice is mostly telling you about last week.

How this was measured

Split-half reliability within a player-season: his average in odd weeks against his average in even weeks, corrected for the halving of sample by the Spearman–Brown formula. Only player-seasons with at least four games in each half count. n = 2,843 player-seasons, 2018–2025, receivers, tight ends and backs.

n=2,843 player-seasons · 2018–2025 · split-half, Spearman–Brown corrected · last verified 22 September 2026

Related studies

Does a new head coach change who gets the ball?It changes WHO, not how muchDoes a receiver struggle in his first season with a new team?His role is less predictableDo touchdowns repeat from year to year?Much less than receptions