Find your playoff edge.

Forecast scores isolate opponent contribution from the offense’s own strength, using only data available at each historical cutoff. Farther weeks use separately fitted horizon effects. Held-out testing supports the current RB and D/ST models. QB, WR, TE and kicker forecast scores remain neutral because their selected candidates did not improve the control. Observed form is available separately; it is descriptive evidence, not a validated forward prediction.

Your overall score is the equal-weight mean of the selected weeks. A selected BYE contributes zero because there is no game to play; it never enters the statistical sample as a zero-point performance. Unknown matchups leave the overall score unavailable. Recency and game averages count appearances, so a bye cannot add a game or age the sample by one played game.

The scores refresh with the existing weekly fantasy pipeline after completed-slate data is ready. Current injuries are shown as context; December availability, late-season rest and future weather are not assumed. Use weekly rankings for start/sit decisions. This score is displayed separately and is not added a second time to their opponent features.

Stats: nflverse; existing D/ST/kicker fantasy scoring; prior-season PFF team grades only where the tested model uses them. No raw PFF feed is redistributed. Research and formulas.