Probetrics — Sports Analytics
NFL slate
Buffalo Bills at Houston Texans
Sun, Sep 13, 1:00 PM ET · Reliant Stadium · Indoors
BUF -1.5 · O/U 44.5
BUF -118 · HOU -102
WinnerProbability the picked team wins, from a normal on the model's predicted margin. Backtested and calibrated: when it says 65%, it has been right about 63–65% of the time.Lean
HOU -102
57%
model margin HOU by 2.5
SpreadWhich side the model's margin lands on versus the spread. A lean, not a proven edge: against closing NFL lines the model covers about 49% in backtest, so this is tracked in public rather than sold.Lean +
HOU +1.5
52%
4.0 pts from the number
TotalModel total (scoring ratings + weather + roof) versus the posted total. Same caveat as the spread: a lean, tracked in public.Lean +
Over 44.5
52%
model total 48.3

Power ratings

TeamEloOffDefSOSGP
Buffalo Bills16044.00.20
Houston Texans16041.82.80

Off/Def are points above league average scored / prevented, opponent-adjusted. SOS is the mean Elo of opponents faced this season.

How the margin was built

Rating gap
BUF +0 Elo
Home field
HOU +2.5 pts
Rest
Even
Weather / roof
Indoors
Model margin
HOU by 2.5
Market spread
BUF -1.5
Model total
48.3 vs 44.5
Moneyline
BUF -118 · HOU -102

Buffalo Bills injuries

  • Tyrell Shavers WROut
  • Zane Durant DTInjured Reserve
  • Jedrick Wills Jr. OTQuestionable
  • Jude Bowry OTQuestionable
  • Keon Coleman WRQuestionable
  • Khalil Shakir WRQuestionable
  • Terrel Bernard LBQuestionable
  • Ty Johnson RBQuestionable

From ESPN's report. Shown for context — the model does not yet price individual absences, which is the known gap between it and the market.

Houston Texans injuries

  • E.J. Speed LBOut
  • M.J. Stewart SOut
  • Ali Gaye DEInjured Reserve
  • Braden Smith OTInjured Reserve
  • Dylan Horton DEInjured Reserve
  • Ja'Marcus Ingram CBInjured Reserve
  • K.C. Ossai LBInjured Reserve
  • Kyonte Hamilton DTInjured Reserve
  • Sam Hagen GInjured Reserve
  • Solomon Byrd DEInjured Reserve
  • Tank Dell WRInjured Reserve
  • Brevin Jordan TEQuestionable

From ESPN's report. Shown for context — the model does not yet price individual absences, which is the known gap between it and the market.

Informational only. Winner probabilities are backtested and calibrated; spread and total calls are tracked leans. 21+, 1-800-GAMBLER.