Wide receiver

Target share counts how often the ball comes; air yards count how far it travels when it does. They are different jobs and the same receiver rarely leads both. A name far to the right with nothing above the median is running the deep routes without being the read, a role that scores in bursts and disappears for a month.

The same two ideas, both now expressed as a share of one offense, which is what makes them comparable across teams. A receiver on a pass-happy offense can post big raw air yards and still own a small slice of them. Below the diagonal is a possession role; above it, a receiver his quarterback is willing to throw past the sticks.

WOPR folds both shares into one usage number, and the vertical spread at any given WOPR is the part of a receiver’s scoring his usage does not explain. Whether that gap is worth chasing is a question about next season rather than this one, and the chart below measures it directly.

What carries into next season

Rank correlation between a receiver's season and his next one, across every pair since 2015. The chances a receiver gets come back; what he did with them mostly does not.

• Targets per gameopportunity0.62
• PPR points per gameoutput0.59
• Air yards shareopportunity0.53
• Target shareopportunity0.49
• WOPRopportunity0.49
• Catch rateefficiency0.43
• Yards per targetefficiency0.25
• Touchdowns per targetefficiency0.17

Spearman rank correlation between a receiver’s season and his next, across 571 paired seasons from 2015 to 2025, each requiring eight games and forty targets in both years. Rank rather than value, because a handful of enormous target-share seasons would otherwise drag the whole coefficient around.

The chances repeat. What a receiver did with them does not. Every measure of opportunity clusters around a half, while touchdowns per target, the thing that most inflates a season line, comes back at 0.17, which is close enough to nothing that last year’s touchdown total should not be read as a forecast. Yards per target is barely better. That is the case for drafting the role rather than the production it happened to yield.

What this does not show. Points per game repeats about as well as the usage underneath it, 0.59 against 0.49 for target share. So this is not evidence that usage forecasts better than scoring does, and the page will not claim it is. It is evidence about efficiency, which is a narrower and more useful thing: the part of a season that came from a high catch rate or a run of end-zone looks is the part least likely to come back.

Boom weeks against bust weeks

Every chart above this one is a season average, and an average cannot tell two very different receivers apart. Keenan Allen and Ladd McConkey finished 2025 within a point of each other per game. Allen was under five points in 12% of his weeks and over twenty in 6%; McConkey was 31% and 19%. Identical on every other chart here, and not remotely the same player to own.

The gap between a receiver’s mean and his median is the same point made a second way. Michael Wilson averaged 13.0 points a game and his median week was 8.4. The average is real; it is just not what most of his Sundays looked like.

Check the box

The operator’s own shortlists, one column per factor. The columns are independent, with different names and different lengths, so a row across the grid means nothing and is not meant to.

Three more columns, fitted rather than judged

The grid above is the operator’s reading of 2025. These three are the same question asked of the play-by-play, and they exist because of the stickiness bars further up the page: touchdowns per target come back year over year at 0.17, but the looks that produce them are opportunity, which comes back at about a half. So the forecastable version of “who scored” is “who was thrown to inside the ten”.

Expected touchdowns are fitted across all 88 qualifying receivers by field zone rather than assumed: an end-zone target is worth 0.35 touchdowns, a red-zone target outside the end zone 0.23, and everything else 0.012, which is most of why a long touchdown is the least repeatable kind. The fit explains 64% of the variance in scoring. The residual is the rest: the part a receiver’s field position does not account for. 49 of the 88 appear on a chart on this page and are the only ones ranked here.