Less noise. A clearer read on the matchup.
Model projections, readable picks, and the thinking behind them.
Washington at New York Giants · November 12. The next edition will compare the spread, total and moneyline after fresh prices, injuries and weather are checked. No new Thursday selection is published here yet.
Sunday’s original spotlight cards remain below as an archive. Historical model differences are estimates, not evidence of a proven betting edge.
Public research tracking begins November 8, 2026. All tracked spotlight picks count. Historical backtests are separate from this forward record.
Original picks and frozen prices.
See Performance for verified settlements.
Giants at Eagles
The model sees a wider scoring gap than the three-point reference spread.
Texans at Chargers
Houston projects to win by about nine. A six-point win returns your stake.
Buccaneers at Bears
The projected total sits about five points below the 47.5 reference line.
Stars indicate editorial research priority, not win certainty or stake size. This prototype has not established a betting edge. Prices are public references, including Canadian listings; verify your sportsbook’s offer.
14 archived matchup projections.
Free with your email signup.
Unlock the original 14 matchup projections and downloadable numbers. Join the pilot list for research, product news, and launch offers.
No card. No automatic billing. Email delivery is coming soon.Already on the list? Enter the same email to unlock this browser again.
Frozen v0.3.1 snapshot · November 7, 2026. Not live odds. The blurred rows illustrate the layout; actual full-slate data loads after signup. Reference prices vary by region. This model has not established a betting edge.
Useful research should explain itself.
Here’s the short version of our process.
Recent scoring, efficiency, pace and opponent profiles establish a scoring forecast. Older games carry less weight.
Weather and provisional injury adjustments modify that forecast. Uncertain lineups get separate scenarios.
100,000 score simulations per matchup show possible results. This is a score model, not a play-by-play game engine.
We compare forecasts with tagged reference prices and explain the gap. A difference is a research signal, not proof of profit.
Our historical checks have not shown a betting edge. Injury point values are assumptions, not validated player effects. The score bootstrap approximates NFL scoring and overtime; historical weather uses recorded conditions, not archived forecasts. Thirteen listed players have unresolved snap matches. Official game-day inactives are pending. More simulations reduce sampling noise, not model error.