Projects each player's passing, rushing, or receiving yards from their own walk-forward season-to-date average (never using games not yet played), then runs it through the same de-vig + trust-guardrail pipeline as the moneyline scanner against live player-prop lines. Receptions, TDs, and anytime-TD props aren't modeled yet — see the note below.
Watch out: This uses a plain season-to-date average, not an opponent-adjusted projection (that's a planned refinement). Validated separately per category on real NFL history — all three beat a naive league-average baseline, but by different margins: rushing cuts error ~34% (correlation 0.59), receiving ~14% (correlation 0.41), passing ~10% (correlation 0.38). Weight a passing or receiving edge with more caution than a rushing one. None of the three know if this week's opponent is a tough or easy matchup yet. Treat flagged picks (a huge gap from the market) as the model probably being wrong, not free money — same rule as everywhere else in this app. Yardage is also right-skewed (a few big games), so the over/under split isn't perfectly 50/50 even when the model is exactly right on average — a real, disclosed imprecision.