The Shrinking Pie

What changed heading into 2026 — a decade of shifts in how fantasy value moves around the field, 2016 to 2025.

Figures computed from xfpgm.com data and the underlying play-level records.

1. The pie is shrinking

Plays per game and total offensive expected fantasy points (XFP) per team both trended down since 2016, bottoming in 2025. The four position columns are the skill-position pie and sum to Total XFP/g; QB passing is excluded (it's the QB's own production, not part of the skill pie). WR lost the most in absolute terms; TE alone held flat.

2016→2025 change: total XFP/g -7.9%, plays/g -4.4%. By position: WR -14.4% (the brunt), TE -1.3% (flat, so TE gained share), RB -7.3%, QB rush +48%. True pass rate fell nearly twice as much as called pass rate (60.1→57.0%, -3.1pp vs -1.7pp): QBs scramble more, so the scramble gap widened 1.8→3.2pp and actual designed-throw volume shrank faster than "pass rate" implies. The 2024 dynamic kickoff lifted average starting field position ~+2 yards (28.5→30.5) but red-zone points/visit and total XFP stayed flat-to-down — the scoring tailwind was absorbed by the scheme drag.

2. Where the pie went: TE is the only real gainer

Split every receiving target's XFP by position — no qualification filter, straight from play-level data.

Complete/unfiltered: play-level rec_xfp by receiver position, every REG target. TE 24.2% in 2025 is a decade high (matches the external record 23.8% TE target share). WR ceded ~2 points in 2025, almost entirely to TE; RB receiving share fell across the decade (18%→15%), it did not rise.

How each position is targeted (depth + per-target value)

TEs are getting more targets but shorter ones — TE aDOT fell from 8.8 to ~6.8 while their share rose. WRs' aDOT and per-target value held. So the volume TE gained is underneath/checkdown work, not deep looks stolen from WRs.

Top-48 WR / top-24 TE / top-30 RB by FPPG each year. The WR decline is a volume story, not an aDOT one: WR aDOT and per-target value held while targets/game fell (7.9→7.4), and the lost volume went to TE, not RB.

3. WR compression — the whole field flattened

The WR position didn't bifurcate into haves and have-nots; it compressed. Rank WRs league-wide by receiving XFP into bands of five, and each band's share of its team's receiving XFP held constant across the decade (top band ~30%, down to ~17% at the WR36-40 band — flat in every year). But the receiving pie shrank underneath those shares. Constant shares on a smaller pie is simple arithmetic: every band falls, and the absolute gaps between bands narrow. That is compression, and it shows up directly in expected points.

WR receiving XFP/game by rank band. Every band fell, the top fell most in absolute terms (−0.79 vs −0.41 at the bottom), and the top-to-bottom gap narrowed 6.35 → 5.98. The elite edge over a startable replacement compressed too (VORP of the 1-5 band over the last startable band fell). In fantasy points, the WR field is flatter and lower, top to bottom — more replaceable, not more top-heavy.

The twist: targets diverged, but efficiency hid it

In targets the elite did pull away — the top band's share of team targets rose while the middle's fell. But that never reached the scoreboard, because the middle's XFP per target rose to offset it (fewer, more valuable looks). So target concentration and XFP compression are both true at once, and the currency that scores — expected points — compressed.

The elite (1-5) grew target share (28.4→30.8%), which held its volume as the pie shrank. The middle (6-10) lost target share (25.5→23.3%) and ate the smaller pie, so its targets fell ~15% — but its XFP/target jumped 12% (1.40→1.57), keeping its XFP share flat. The raw-target "gap widening" is real; it just gets neutralized by efficiency before it becomes fantasy points.

It's not the alpha getting devalued either

And it isn't the within-team WR hierarchy flanderizing. The WR1's share of his team's total receiving XFP is flat at ~23% for a decade, the count of dominant alphas (WR1 ≥28% of team rec XFP) shows no downtrend, and the WR1/WR2 ratio is flat-to-up (1.48→1.57) — they're not converging. The within-team hierarchy is unchanged; the compression comes from the shrinking pie, not from any reshuffle of shares within a team. Same story among WRs specifically: the lead WR still takes ~38% of his team's WR receiving XFP and the top two ~64%, unchanged.

4. Tight end — rising share, deeper startable field

TE18 FP jumped to 8.3 (from ~6.5) — the deepest startable TE field of the decade — and TE1 XFP (opportunity) is a decade high. Note TE1 FP (14.9) sits well below the Kelce-apex years (17.4 in 2020): the position shifted from one dominant scorer to a deeper young cohort.

5. Running back — more share, lower ceiling

The workhorse "revival" is a usage story, not a production one. Lead backs command a bigger slice of the run game, but it's a smaller run game — so the high-end RB ceiling actually fell. RB1 total (rush+rec) FP/g dropped from ~25-27 in 2016-2020 to ~21.5 now.

Top RB totals (rush+rec) per game, deduped. Realized fantasy points at the very top fell — RB1 ~24-27 (2016-20) to ~21.5 now — but that is TD regression, not lost opportunity: RB1 XFP barely moved (22.8→21.6). The clearer picture is by rank band below.

RB opportunity concentrated to the top — the mirror image of WR

Rank RBs league-wide by total (rush+rec) XFP into bands of five, through RB40. Unlike WR, the RB field did not flatten: the top bands held or rose while the depth (bands 26-40) fell, so the gap widened. Touches concentrated onto fewer elite workhorses.

RB total (rush+rec) XFP/game by rank band. The top (1-5) held and even rose (16.28→16.69) while the depth fell (36-40: 7.23→6.77), so the top-to-40 gap widened 9.05 → 9.92 — the exact opposite of WR, which compressed. This is the bellcow concentration in opportunity terms: fewer, bigger workhorses over a thinner committee tail. It's why elite RB opportunity is worth reaching for even as the realized ceiling regressed. Now the usage side, where the "back is back" claim holds directly:

Lead-back rush share is at the decade high (67-68% in 2024-25), true bellcows near 19, and 3-down dual-threats up from 10-13 mid-decade to 17-20. Externally, six RBs logged 300+ carries in 2024, and RB receiving TDs crossed 100 in 2025 for the first time since 2021.

6. Quarterback — passing fades, legs are the bankable floor

QB fantasy value is migrating from the pocket to the ground. Per starting QB, passing XFP fell from 15.3 to 13.3 per game while rushing XFP rose — so rushing's share of a QB's expected fantasy points climbed from 11% to 16% (peaking ~18% in 2023).

Per starting QB (≥150 season pass XFP). The two inputs behave completely differently year to year: rushing XFP is a near-locked trait — same-QB persistence r=0.90, while passing XFP is noisy (r=0.47). A running QB is a floor you can bank on draft day; a pocket passer's weekly output swings.

The tiers: a high floor, and streaming is cheap

Rank qualified starters by total (pass+rush) XFP into bands of five, QB1 through QB20. The whole position drifted down, and the top-to-bottom gap widened only slightly — the rushing QBs at the top held their floor while pocket-passer-only QBs at the bottom faded with the passing game.

QB total (pass+rush) by rank band, qualified starters (~22-27/yr). Two things moved at once. (1) The elite came down: the dual-threats toned down their passing — Allen 18.1→11.3, Mahomes 17.4→12.7, Lamar 12.1→8.2 pass XFP/g — and some of the running too (Lamar 6.8→2.9, Hurts 8.3→5.3 rush XFP/g in 2025). So the 1-5 band's value-add shrank. (2) But the 11-15 tier fell further: with no rushing floor, pocket-only mid QBs can't make up ground as passing compresses, so the gap from elite to 11-15 actually widened — XFP 3.7→4.3, actual FP 4.9→5.5. The 11-15 band sank toward replacement (15.9→14.2 FP) and is now a dud zone.

Passing efficiency over expected (OEX) is a real skill but only loosely repeatable (year-over-year r=0.23; top-decile passers convert +21% over expected, bottom −15%). Two archetypes fall out — and they are exactly what the streaming pair is built from:

Lamar Jackson is the outlier on both lists (7.7 rush XFP/g and a 1.15 passing OEX) — the rare QB who is his own floor and ceiling. For everyone else the edge is the pairing: a cheap rushing QB banks the floor every week (the stable input), a proven passer supplies the matchup-driven ceiling (the volatile-but-skilled input), and you start whichever the week favors.

7. The scheme shift — proof

Defenses went two-high; offenses answered with more two-TE (12-personnel), routing underneath targets to tight ends. Coverage is charted 2018+.

Motion, RPO, and QB legs all rose (motion charted 2022+, RPO/PA 2019+):

Two-high (Cover 2/4/6/2-Man) climbed 32.9%→43.0%. Motion rose 33%→54% (matching PFF's ~64% on their fuller series). RPO rose then plateaued; play-action is choppy. QB scrambles/game climbed steadily 1.15→1.92 — QB legs are a growing share of the position's value. Motion rose smoothly with no dip after the 2023 motion-loophole rule change, so there's no discontinuity to control for.

8. How we trade on it

Sources & method

Internal: team tendencies/splits, team xfp_summary, play-level xfp_play_level (complete receiving XFP by position; coverage type), qb_rush_metrics (scrambles), and player scatter files (qualified receivers ≥30 targets / backs ≥20 carries). Concentration uses rank cohorts per year; RB counts filter adj_gp ≥ 8. FPPG is full-PPR unless noted. Two-high = Cover 2/4/6/2-Man from FTN coverage charting (2018+).

• CBS — Dave Richard, 2026 WR tiers (WR compression, TE share/TDs)
• Yahoo — Matt Harmon, 2025 WR Exit Interview (RB-over-WR flex)
• PFF — Pre-snap motion usage reaches new highs
• SI — Passing-game decline, two-high record
• Colts.com — Workhorse RBs comeback 2024
• NFL Operations — Dynamic Kickoff second season