- Sleeper
- Sleeper player projections rolled into a team score: 6 × (rush TD + rec TD + def TD) + 3 × FG + XP. Passing TDs are never counted — that would double-count the same score. Votes both the spread and the total.
- FantasyPros
- Same scoring as Sleeper, different projection book. Weekly consensus stats from the FantasyPros partners feed (QB / RB / WR / TE / K / DST). If Sleeper and FP disagree, each retains its own vote; five of seven must agree. Votes both the spread and the total.
- Form
- Predicted scoreline from recent results: 60% last five games scoring / allowed, 40% last three, then blended with the opponent (your offense vs their defense). Home field adds 1.25 points to home and subtracts 1.25 from away. The spread margin then shifts 1 point toward away; total unchanged. Needs four games of history.
- Elo
- Homemade franchise ratings, not FiveThirtyEight. Every team started 2025 at 1500; those finals were replayed into the seed you see in week 1. Predicted home margin is (home − away) / 25 + 2.5— 25 Elo is about one point, and home field is 2.5 points (62.5 Elo). After you grade a week, ratings move by K = 20 × |margin|^0.25. Elo votes the spread only. It does not know injuries, rest, or totals.
- Pace
- Last-five combined scoring — how many points have gone up in these two teams’ recent games, not who should win this matchup. Formula: (home PF+PA + away PF+PA) / 2. No opponent blend, no last-three tilt, no home field. That’s the difference from Form O/U: Form builds a scoreline (offense vs defense); Pace only asks whether this game lives in a high- or low-scoring environment. Votes the total only.
- EPA
- One of seven votes on both the spread and the total. nflverse already computed EPA on every play — we roll 2025 regular-season pass/run snaps (plus any completed 2026 weeks) into team rates. Offense and defense EPA/play are shrunk toward the league mean (n₀ = 200 snaps), then blended 60% season-to-date / 40% last five games. Matchup EPA/play is (your off + their def) / 2, times expected snaps (pace_home + pace_away) / 2, added to league-average points with +1.8 home field. Week 1 uses 2025 only so there is no leakage from this week’s games.
- Weighted history
- Last 17 games with a five-game half-life. Unshifted vote on spread and total.
- Possessions
- Expected possessions and opponent-adjusted points per drive. Spread margin shifts 1 point toward home; total unchanged. Missing inputs PASS.
A play prints only when five of seven indicators agree. Close games PASS: 2 points on the spread, 2.5 on the total. Six agreeing is green; seven is gold. Totals are a work in progress. Original and active predictions are both saved for comparison. After a play prints, shop FanDuel vs DraftKings: the extra half-point on the spread or total always beats better juice. Same number → both chips. On the full slate, tap an indicator to see the players, last five games, Elo, or EPA math.