Projections with receipts.
Every ranking on this site comes from a model that was backtested before it was believed — and every method is published in full. No film takes, no vibes, no hedging.
Draft 2026 season board
A ridge-regression projection over 8 seasons of data — usage, TD luck, age, draft capital, salary structure. It out-ranks the ADP consensus at QB, WR and TE, and says so with confidence intervals.
+.08 Spearman vs ADPWeekly Full-position boards
Skill positions from a validated weekly model; D/ST and kicker streaming straight from the one number that matters — the Vegas line — plus the few tiebreaks that survived testing.
QB .43 · RB .63 · D/ST .30 rank corrResearch The findings
28 numbered findings with methodology and honest caveats: what actually wins drafts, which folklore survives testing, and which popular stats are noise.
28 findings · 6 simulated seasonsLive Draft assistant
The full-screen draft-day tool: live board, value alerts, pick sync. Opens in distraction-free mode — you'll want the whole screen.
launches full screen →Weekly boards
Ranked by opponent implied total — worth .304 within-week rank correlation on its own; everything else anyone tells you about D/ST streaming adds ≤.005 (finding 27).
Ranked by own-team implied total with a dome bump. The kicker himself barely matters — stream the situation (finding 28).
Skill-position boards go live with the season
The weekly model needs two weeks of current-season form (it blends trailing production with usage-based expected points, opponent strength and the Vegas total). First boards post September 2026 — with movement reasons.
2026 season board
returning players · model v1 · rookies via ADP until the rookie model shipsMarket environment — what Vegas already thinks of every 2026 offense
Implied points per game across all 272 games of the 2026 schedule (consensus of ~9 books, snapshot Aug 2026). Use it as draft-day context: a top-5 environment lifts every pass-catcher in it.