The average WR week overstates the typical one: rank 2025's top 36 by median, not mean, and 16 move five-plus spots

A drafter reads one number per receiver — the season average, or the total it sums to — and sets a lineup expecting roughly that most weeks. We scored all 18 of the 2025 regular season's weeks in full-point PPR from the per-player weekly rows of nflverse's stats_player_week_2025 file, then took every wide receiver who played at least 14 of those 18 games — 94 of them, the complete set in that file. For 78 of the 94, the per-game average sat above the median week — the game they actually turned in half the time — because fantasy scoring is right-skewed: a few big afternoons pull the mean up past the middle. Rank the 36 highest by average (WR1–WR36, the weekly-startable pool in a 12-team, three-receiver league) and 28 of those 36 still carry an average above their median. Re-rank the same 36 by median week instead — the typical game, not the average — and 16 of them move five spots or more. The widest gap is Arizona's Michael Wilson: WR13 by average (13.0 per game), WR35 by median (8.4), a 22-seat fall, with his three biggest weeks (37.2, 33.5, 21.8) accounting for 42% of his season. A.J. Brown ran the same shape one tier up — WR9 by average (14.7), WR17 by median (12.1), 40% of his year in three games (35.2, 28.1, 25.0). At the other end sits the metronome: Puka Nacua, the overall WR1, averaged 23.4 and posted a 23.0 median — a 46.5-point ceiling and a 4.8-point floor washing out to nearly the same number.

The following is analysis, not fact. A season average is a sum divided, and division hides the shape of what it divided. The single-team manager who starts one receiver each week rarely collects exactly the average; they collect a specific game, and across a season the median can describe that game better than the mean when a player's scoring is skewed — which, in this pool, it usually is ^mmodeled. The wedge between the two rankings is boom concentration: the receivers who fall when you switch to median are the ones whose season number leans on a handful of big weeks; the ones who hold — or climb — carry a median at or near their mean, a shape consistent with production spread across more weeks rather than piled into a few ^mmodeled. For eight of the top 36 the median actually sat at or above the average; Indianapolis's Michael Pittman (11.9 average, 13.7 median) rose from WR21 to WR9 on that median-mean gap. A median above the mean marks the direction of the skew, not the absence of bad weeks — Pittman still had several single-digit weeks. None of this says a spike-driven line repeats or regresses — whether 2025's skew carries into 2026 is a separate question this record cannot answer, and any such expectation is a projection, not a reading ^mmodeled. What the record does show is that two receivers can carry the same average and hand a lineup very different weeks.

The median is not a verdict. A receiver whose value arrives in bursts loses nothing to timing in a season-long points-total format, where only the sum matters, not when the points land; best ball rewards the spikes further, counting each week's best lineup so a boom banks in full while a dud can sit on the bench — and if his booms are even loosely predictable, the spikes are the point, not a flaw ^mmodeled. The median has its own blind spot: it cannot see when the weeks fell, so a receiver who front-loaded a strong first half and faded can post a flattering median off a collapsed second half, which is part of why a name like Pittman reads steady on this measure despite a stretch of single-digit weeks. And this is one season in one scoring format: full PPR flatters high-volume possession receivers, so a standard or half-PPR ledger would reshuffle which names read as spike-driven versus steady — the skew is general, the exact ordering is not. A manager who treats the season average as the signal and the weekly scatter as noise is making a defensible call the data does not settle.

From a single roster, a receiver is last week's box score and a season-long number on a draft page. From the whole pool — all 94 who played 14-plus games, all 18 of everyone's weeks scored at once — the number and the game it summarizes come apart, and the record shows exactly how far apart, receiver by receiver.