Performance leaders

The default performance rating combines 70% fantasy output, 15% net yards per opportunity and 15% EPA per opportunity. Production is scaled by the actual gap between the lowest and highest output in the rated position group (equal outputs score 50). Efficiency uses position-relative percentiles, moderated toward 50 by measured opportunities / (measured opportunities + the position’s minimum). If fewer than two reference players qualify, efficiency is neutral at 50. Rates are calculated over the whole window, not averaged from weekly rates. Total EPA is displayed for context but does not receive an additional weight.

QB efficiency uses dropbacks, including sacks and scrambles; RB uses carries plus targets; WR/TE use targets. All fantasy production still counts. No minimum workload excludes a high scorer from the composite if complete EPA and yardage exist. Missing measurements withhold a rating rather than inventing one; Fantasy production still shows their known points. The lower-rated list requires the box-score workload minima shown above. These relative ratings are not opponent-adjusted, predictive, isolated player skill, or comparable between positions or weeks. A different cohort or scoring format can change the rating. The 70/15/15 weights are an explicit product policy, not scientifically fitted weights.

The optional Fantasy production view: order by recorded fantasy points in the selected format, or points per opportunity game for the optional season view. There are no EPA weights in this ranking. Missing EPA does not remove known fantasy points. The lowest scoring list uses box-score workload minima and excludes the displayed top group. Equal point values share a rank; stable player ID decides order at the five-player boundary. A minimum workload is not a verified start, and lower production is not necessarily a disappointment versus expectations.

Opportunity games count at least one recorded pass attempt, carry or target, with all those counts known throughout the window. A zero-opportunity appearance is not counted; this measure must not be read as official games played. Season totals do not require two games; the per-opportunity-game view and lower-scoring list do. Missing fantasy results exclude an incomplete total rather than treating it as zero.

Experimental index only

Each metric is ranked within the qualified position group. A percentile of 100 is highest and 0 is lowest; equal values share the midpoint of their places. The score weights these percentiles: 50% fantasy-point total, 25% total EPA, 25% EPA per opportunity. The season per-game view replaces fantasy and EPA totals with each player’s values per recorded role game before ranking. Efficiency always uses total EPA divided by measured opportunities. These fixed weights define a descriptive performance score; they have not been fitted to predict next week.

QB: dropbacks (including sacks and scrambles) plus designed carries. RB: carries plus targets. WR and TE: targets. EPA means expected points added by the recorded play, which is shared with teammates; it does not isolate individual credit or adjust for opponent strength. The denominator counts only the eligible plays in these metric definitions, so it can differ from box-score attempts or targets.

Weekly totals use one completed week. Season totals accumulate Weeks 1 through the selected week. Season EPA per opportunity divides total EPA by total eligible opportunities; it does not average weekly rates. Fantasy totals are summed at their recorded two-decimal precision.

Weekly minima: QB 15 dropbacks, RB 8 carries plus targets, WR 5 targets, TE 3 targets. Season requires at least two games with recorded role opportunities and the same minimum workload per such game. These are workload requirements, not verified starts or snap counts. Inactive zero-filled rows do not count as games played. Missing required measurements exclude a player; weights are never silently reassigned.

Scores use full precision before display rounding. True ties share a position rank, with player ID deciding list order at the five-player boundary. At least ten qualified players and a difference in scores are required to show the two lists. Scores compare players within one position and window, not across positions or weeks.

QB, RB, WR and TE are supported. K and D/ST do not yet have comparable published advanced-player measurements. These are completed-performance rankings; use the weekly projections for the upcoming slate.

The latest completed week follows the published data. This page refreshes on focus and every five minutes while open; new accepted weekly data updates the rankings without a code release.