NFL fantasy usage trends

The three numbers that matter

You don't need a dozen advanced stats. Three cover most of what determines a player's fantasy ceiling and floor:

1. Snap share. The floor underneath everything else. A player can't produce from the sideline. Watch for a jump of 15 points or more in snap share across back-to-back weeks. That's usually a real change in the coaching staff's plan, not noise.
2. Target share (or carry share for backs). The single best predictor of future receiving or rushing production. A player getting 25% of his team's targets is a weekly starter in most formats regardless of last week's box score.
3. Red zone share. Touches inside the 20-yard line convert to touchdowns far more often than touches at midfield. A player who becomes his team's primary red zone option gains touchdown equity that doesn't always show up yet in his yardage totals.

Why the rolling window matters

A single game is a small sample and it lies to you. A receiver can see 12 targets in one game because the opponent had two defensive backs injured, and go back to 6 targets the next week once the matchup normalizes. A 3-game rolling window smooths that out and shows you the underlying role instead of one week's noise. A 5-game window is slower to react but more reliable for identifying a real change versus a blip.

The trade-off: a shorter window reacts faster to real changes (an injury ahead of a player, a new offensive coordinator) but also reacts faster to noise. Use the 3-game number to catch something early and the 5-game number to confirm it before you act on it with a waiver claim or a lineup decision.

Role versus performance

Keep these separate. Role is opportunity: snaps, targets, carries, red zone touches. It's largely under the coaching staff's control and it's sticky from week to week. Performance is what a player does with that opportunity: yards per target, touchdown rate, missed tackles forced. Performance regresses toward a player's established level far more than role does. When you're projecting a player forward, weight role heavily and treat a hot or cold performance streak as more likely to normalize than continue.

One confirmed data point beats one outlier

The single most common mistake in reading trends: reacting to one big game. Wait for a second data point in the same direction before you move a waiver bid or a start/sit call on it. If the role change is real, it shows up again the following week. If it doesn't, you saved your budget for something that actually holds.