How Accurate Are Fantasy Football Projections? What Players to Trust and What to Ignore
Projections are estimates, not predictions. The distinction matters because it shapes how you use them. A projection tells you what a player is expected to average across many games in similar circumstances — it does not tell you what will happen in a specific game.
Projection Reliability Quick Reference
| Player type | Projection reliability | Why | How to use |
|---|---|---|---|
| High-volume WR with consistent role (9+ targets/game) | High — volume drives accuracy | Volume is the most stable input; consistent target share = consistent scoring | Trust as a primary signal; projection reflects real opportunity data |
| Committee RB (50-50 split) | Low — volume unpredictable week-to-week | Game script, health, and coaching decisions shift carries | Use projection only directionally; other inputs (matchup, opponent run defense) matter more |
| Elite QB in high-passing-volume offense | High for season average; moderate for single week | Season average well-predicted; single-game variance from interceptions, INTs, game script | Trust for who to start; do not trust for single-game point total |
| Red zone touchdown projection for any player | Very low — TDs are the highest-variance category | 0.6 TDs/game projected means almost anything can happen in any specific game | Ignore TD projection for week-to-week; only use for season-long valuation |
| Player returning from injury with limited clarity | Very low — probability of playing and snap limit both uncertain | Projection often built on “average” of healthy/unhealthy scenarios | Weight injury report and practice reports far more than the projection number |
| Two players projected within 2–3 points of each other | Essentially equal — projection cannot differentiate | Projection error is often larger than the gap between them | Do not use projection alone; use matchup, recent form, and usage trends to decide |
What Projections Actually Represent
Projections are typically generated by combining several underlying estimates:
- Target or carry volume: How many opportunities is this player expected to receive?
- Success rate: At what efficiency per opportunity does this player typically convert?
- Touchdown probability: Based on red zone usage and opportunity, what is the expected TD contribution?
- Scoring format adjustment: PPR adds points per reception, which changes WR and pass-catching RB projections relative to standard.
Each of these underlying estimates carries uncertainty. When you combine three or four uncertain inputs, the combined output has more uncertainty than any single input alone. This is why the best projections are presented as ranges, not single numbers — a player projected for “12.4 points” is really being projected for “somewhere in the 6–18 range, most likely around 12.”
Where Projections Are More vs Less Accurate
More reliable:
- High-volume players with consistent roles — a WR who runs 90% of routes and sees 10 targets per game has a more predictable projection than a committee RB
- Players with stable offensive situations — a QB in a high-volume passing offense who has run the same scheme for 2–3 years has a tighter projection range
- Season-long averages — projections are more accurate as season-length estimates than week-to-week predictions; over 17 games, variance smooths out
Less reliable:
- Running backs in committees — volume depends on game script, health, and coaching decisions that are genuinely hard to predict
- Red zone touchdown projections — touchdowns are the most volatile fantasy scoring category; a WR projected for 0.7 TDs per game might score 0 or 2 in a specific week with no meaningful difference in opportunity
- Players in new situations — new team, new coordinator, new role each adds uncertainty that projections cannot fully capture
- Players returning from injury — a “questionable” tag means probability of not playing, probability of limited snaps, and probability of full workload; projections handle this crudely at best
The Projection Error Range
In practice, individual game projections are wrong by a significant margin the majority of the time. Studies of projection accuracy across platforms typically show that:
- The correlation between a player’s projection and their actual score is meaningful but not strong enough to use as a primary decision driver
- Projection error for a single player in a single game is often as large as the projected value itself
- Rankings based on projections have genuine predictive value when comparing players at the extremes (high-projection players outperform low-projection players on average), but the ordering among similarly projected players is close to random
How to Actually Use Projections
Use projections for coarse decisions. If one player is projected for 18 points and another for 9, the 18-point projection reflects real differences in role, volume, and opportunity that should inform your decision. Trust the coarse signal.
Ignore projections for fine decisions. If both players are projected for 13–14 points, the projection is not differentiating them meaningfully. Use other factors: injury reports, historical performance in similar matchups, recent usage trends.
Use projections as a cross-check. If your gut says Player A is the right start and the projection says Player B, it is worth asking why the projection differs from your instinct. Sometimes it reveals something you missed. Sometimes your reason is correct and the projection missed something.
Never rely on a single projection source. Different projection systems make different assumptions. Averaging projections from multiple sources reduces the impact of any single system’s bias or error.
For how to evaluate matchups that the projections may not fully capture, see: Fantasy Football Matchup Analysis: How to Use Defensive Matchups to Make Better Start-Sit Decisions. For how target share and snap count data provide a more grounded view of player opportunity, see: Fantasy Football Opportunity Share: How to Evaluate Rush Share and Target Share.