What STRONG and BET actually mean
Every NFL pick is filed in one of two tiers based on a strict combination of model edge, EV score, and situational compounding. This page documents exactly what each tier requires, what historical win rate to expect, and which markets we never publish.
Edge ≥ 12 yds · EV ≥ 0.55 · Situation score ≥ 2
For a yards pick to be tagged STRONG, the model has to find a 12+ yard gap between projection and bookmaker line, an EV (edge/volatility) score of at least 0.55, AND at least 2 situation factors confirming the read (defence quintile, total, dome, spread, week-of-season, etc.).
TE picks raise the edge floor to 15 yds. Weeks 1-3 apply a 1.30× threshold multiplier (so 15.6 yds and EV 0.72 in early-season).
Edge ≥ 7 yds · EV ≥ 0.40 · Situation score ≥ 1
BET tier covers smaller-edge picks that still pass the calibration filter. Use these as supplementary rather than core positions. The salvage rule allows situation-verified BETs: combined edge score ≥ 3 with EV ≥ 0.45 qualifies even with smaller raw edge.
QB passing is never published at BET tier — σ = 58 yards is too wide for edges below 0.55 EV. We hold QB exposure for STRONG-only.
What we never publish
- TE picks with edge below 15 yds (34% historical win rate at default 12yd floor)
- QB passing picks below STRONG tier (σ too wide for low-EV reads)
- Receiving / passing picks when wind ≥ 20 mph regardless of edge
- Anything where games_played < 4 in the new season (small-prior dominance)
- Playoff QB rushing — model has -4.7 yd documented bias
Weekly pick caps
To prevent overcommitment to a single slate, we cap selections by market:
| Market | Max picks/week | Notes |
|---|---|---|
| Receiving yards | 4 | Max 1 TE |
| Rushing yards | 2 | — |
| Passing yards | 2 | STRONG only |
| Anytime TD | 3 | — |