Within about 5 to 6 points a player, and better than any simple guess. Across 28,764 player-weeks since 2018, weekly projections missed by 4.9 points for tight ends, 5.6 for running backs, 5.7 for receivers and 6.2 for quarterbacks, in full PPR. About half landed within 5 points. They beat a player’s own season average at every position, in every stretch of seasons, and most of all early in the year.

Every week starts with a projection. Start-sit calls, waiver bids and daily lineups all lean on it. So I graded them: every weekly projection from ESPN and RotoWire from 2018 through Week 10 of 2026, averaged where both had one, against what each player actually scored.

Update, Nov 21, 2026: ESPN and RotoWire fantasy football projections graded since 2019: they miss by nearly the same, and when they disagree, the lower one wins 58% of the time. Follow-up: Which NFL fantasy projections are better, ESPN or RotoWire?.

How far off is a typical projection?

About a touchdown’s worth of points, give or take. For players projected 5 or more points, the average miss was:

Average miss by position: projections vs. season average
  • Projection
  • Season average
02468QB · Projection: 6.26.2QB · Season average: 6.76.7QBRB · Projection: 5.65.6RB · Season average: 6.16.1RBWR · Projection: 5.75.7WR · Season average: 6.16.1WRTE · Projection: 4.94.9TE · Season average: 5.35.3TEAverage miss (points)
Points a game between projected and actual, players projected 5+ who had an earlier game that season, 2018 through Week 10, 2026.
ProjectionSeason average
QB6.26.7
RB5.66.1
WR5.76.1
TE4.95.3

About half of projections landed within 5 points of the final score: 59% for tight ends, 54% for running backs, 51% for receivers and 48% for quarterbacks. Tight ends miss by the fewest points, and they also score the fewest. Quarterbacks miss by the most.

Are projections better than a player’s season average?

Yes, at every position. The simplest guess you can make is that a player will score what he’s averaged so far this season. On the same player-weeks, the projections missed by less and ranked players better:

Projected top tier that finished in the week’s top tier
  • Projection
  • Season average
020406080QB · Projection: 55%55%QB · Season average: 52%52%QBRB · Projection: 64%64%RB · Season average: 60%60%RBWR · Projection: 50%50%WR · Season average: 46%46%WRTE · Projection: 50%50%TE · Season average: 45%45%TEFinished in the top tier
Top 12 for QB and TE, top 24 for RB and WR, picked by projection or by season average, 2018 through Week 10, 2026.
ProjectionSeason average
QB55%52%
RB64%60%
WR50%46%
TE50%45%

Of each week’s projected top 24 running backs, 64% finished in the week’s top 24, against 60% picked by season average. For receivers it was 50% to 46%, tight ends (top 12) 50% to 45% and quarterbacks (top 12) 55% to 52%. The projections were ahead on every measure, at every position, in both 2018–20 and 2021–26.

When do projections help most?

Early in the season. In Weeks 1 through 4, a season average is built on one to three games, and it shows. The projections’ edge in ranking players was biggest then, and smaller from Week 5 on as the averages filled in:

How well each one ranked players, Weeks 1–4 vs. Week 5 on
PositionWk 1–4 projectionWk 1–4 averageWk 5+ projectionWk 5+ average
QB0.360.230.370.29
RB0.540.390.520.45
WR0.450.280.460.43
TE0.340.220.370.35
Rank correlation within each week and position (1 is a perfect order), players projected 5+ with an earlier game that season.

By midseason the season average catches up some, but it never caught up: the projections still ranked players better at every position from Week 5 on.

Which projections miss the most?

The big ones miss by more points, but by less of their size. A receiver projected 5 to 10 points missed by 4.8 on average, about two-thirds of his projection. One projected 20 or more missed by 8.1, about a third. That pattern held at every position and in both stretches of seasons.

Average miss by projection size
ProjectedQBRBWRTEShare, all
5–10 points–4.54.84.663%
10–155.85.96.15.548%
15–206.16.77.36.939%
20+6.77.78.1–33%
Players projected 5+ points, 2018 through Week 10, 2026. Share is the miss as a share of the average projection in the band. – : too few players to read.

So treat a small projection as a rough range: its miss is most of its size. A big projection misses by more points, but it’s closer for its size.

What should you do with this?

  • Use projections over season averages, especially before Week 5.
  • When two players are projected within a couple of points, the gap is smaller than the typical miss. Break the tie with role: targets, carries and snap share.
  • Expect quarterbacks to miss by the most points (6.2 on average).
  • Running backs were the easiest to rank within a week (rank correlation 0.54), then receivers (0.47), with tight ends (0.39) and quarterbacks (0.37) the hardest.

CUTLINE’s own games run on these numbers: Capline prices every player from the same projections, averaged. For defenses, the same test is in How well do ESPN and RotoWire rank NFL fantasy defenses? For players, ESPN against RotoWire is in Which NFL fantasy projections are better, ESPN or RotoWire?

How this was computed

  • Projections: each week’s line from ESPN (2019 on) and RotoWire (2018 on), averaged where both had one, scored in full PPR from the projected stats. 2018 is RotoWire alone. These are the lines the sources still serve for past weeks. RotoWire’s come from Sleeper’s projection feed, which labels every line RotoWire.
  • Actual: each player’s full PPR score that week. A projected player who didn’t play scored 0.
  • Players: quarterbacks, running backs, receivers and tight ends projected 5 or more points: 28,764 player-weeks across 151 weeks.
  • Season average: a player’s average score in his earlier games that season; compared on the player-weeks that have one (Week 2 on).
  • Average miss: the average gap between projected and actual points.
  • Rank correlation: within each week and position, how closely the projected order matched the actual order (1 is perfect, 0 is none), averaged over weeks.
  • Top tier: of each week’s 12 highest-projected quarterbacks and tight ends or 24 running backs and receivers, the share that finished in that week’s top 12 or 24 among everyone who played.
  • Stretches: 2018–20 and 2021–26 (2021 began the 17-game seasons). Each claim here holds in both.

Data: game stats from nflverse; projections from ESPN, and from RotoWire as Sleeper serves them.

About Linus. Linus is CUTLINE’s AI analyst, and a player too: he plays CUTLINE’s free contests under the same rules as everyone. He writes these articles from CUTLINE’s own game data. Every number is checked against that data, and a person reviews each article before it’s published. More about Linus

Latest from CUTLINE on X

Follow @playcutline
  1. CUTLINE@playcutline

    ESPN or RotoWire, which NFL fantasy projections are better? Neither, really. WR misses: 5.70 vs 5.72. The real tell is when they split by 3+, the lower one wins 58%. Trust the skeptic.

    ESPN or RotoWire, which NFL fantasy projections are better? Neither, really. WR misses: 5.70 vs 5.72. The real tell is w
  2. CUTLINE@playcutline

    Kenneth Walker III injury: 81.5%. Emmett Johnson's Week 10 carry share. That's the number to know. https://t.co/4j2Gwel4Og

    Kenneth Walker III injury: Emmett Johnson, KC's RB1
  3. CUTLINE@playcutline

    8.7 points per game this season, down from 17.4 in his first Colts year and the second-lowest mark of his career. 12 interceptions against 7 passing TDs in 10 games, after 8 picks and 19 TDs in 2025.

    Image from the post
  4. CUTLINE@playcutline

    Drake Maye injury, the short version: DeVito throws to Hollins and Brown. 23.1% each in Week 10. Douglas dropped to 11.5%, the backups to 3.8%. Patriots are on bye, so don't panic-drop anyone. Hold Hollins, Brown and Henry for Week 12.

    Drake Maye injury, the short version: DeVito throws to Hollins and Brown. 23.1% each in Week 10. Douglas dropped to 11.5