PRO CONTENT · FREE

QB fantasy projections are a coin flip — a well-made one

Sunday Sidekick Research · August 12, 2026 · full-PPR scoring

Here is the most surprising thing our data says about quarterbacks: the projections are not too high and not too low. Across 806 weeks where a QB was projected 15+ full-PPR points, he beat the number 49.9% of the time and the median miss was exactly 0.0. As a midpoint, a QB projection is about as honest as a number can be.

The catch is everything around the midpoint.

Projected	Weeks	Beat proj.	Within ±3	Lost 5+	Gained 10+	Median	Typical range (p10–p90)
3–9 pts	18	28%	67%	11%	0%	-2.3	-5.4 to +0.9
9–15 pts	270	50%	35%	23%	8%	+0.1	-9.2 to +8.7
15+ pts	806	50%	29%	26%	11%	+0.0	-9.2 to +10.4

The math of a "20-point" quarterback

Take a QB projected at 20.0. The middle 80% of comparable weeks landed between the 10th and 90th percentile of misses — that is, between 20 − 9.2 = 10.8 and 20 + 10.4 = 30.4. One week in four (26%), he finishes 5+ points under his number. One week in nine (11%), he beats it by 10 or more. The projection told you the middle of the road, not the width of it — and at quarterback the road is about twenty points wide.

Why 1-point ranking gaps are noise

Only 29% of elite-projected QB weeks land within ±3 points of the projection. The weekly standard deviation is about 7.7 points. When two QBs are projected 19.5 and 21.0, the 1.5-point gap is roughly one fifth of one standard deviation — the "wrong" QB outscores the "right" one constantly, and that is not the ranking being bad, it is the position being wide.

What we do with this: treat a QB projection as a fair midpoint, never a promise. Between two options separated by a point or two, the projection has no opinion — break the tie on things the single number can't hold, like rushing volume (a floor that throws incompletions can't erase) and the game you expect the offense to be in.

The bust rate is a schedule, not a scandal

A quarter of high-end QB weeks come in 5+ points under projection. Over a 14-week fantasy season, that is three or four ugly weeks from a QB doing exactly what the data says a good QB does. The mistake is reacting to week 3 of 14 as if the distribution owed you its median every Sunday.

Methodology. A residual is a player's actual full-PPR score minus his pregame projection for that week. Sample: 1,094 scored QB player-weeks across the 2024 and 2025 regular seasons (9,254 player-weeks measured in total, all positions), grouped by the size of the projection. Weeks with a projection under 3 points are excluded as deep-bench noise. Projections are a well-known industry consensus feed; the measurement and analysis are ours. If you reference these numbers, please link to this page as the source.

These numbers come to life on Sundays. Sunday Sidekick shows your players' points moving live, play by play, across every league you're in — free. Get the app.