Task Enterprises presentsNFL Intelligence

2013 · Week 3

Final

Rams

St. Louis

QB Sam Bradford

7·31

Cowboys

Dallas

QB Tony Romo

AT&T Stadium · closed · matrixturf

Closing spread

DAL -3.5

Total

47.5

Model margin

+1.8est.

DAL perspective

Push chance

None

half-point line — cannot push

Win probability

  • Cowboys
  • Rams

How each play moved the game.

Cowboys winsRams wins

Matchup

Where the edges are

Season-long league percentiles for both sides, mirrored. The saturated side of a lopsided row holds the edge.

The rivalry

DAL lead 8–5 since 1999

DAL have won 2 straight. Meetings average 46.2 points. DAL are 8-5 ATS in the series.

Previous meetings
SeasonMatchupScoreWinner
2023 · W8LA @ DAL2043DAL
2022 · W5DAL @ LA2210DAL
2020 · W1DAL @ LA1720RAMS
2019 · W15LA @ DAL2144DAL
2018 · playoffsDAL @ LA2230RAMS
2017 · W4LA @ DAL3530RAMS
2014 · W3DAL @ STL3431DAL
2011 · W7STL @ DAL734DAL

Box score

Individual performances

Cowboys

Passing

PlayerC/AttYdsTDINT
Tony Romo17/2421030

Rushing

PlayerCarYdsAvgTD
DeMarco Murray261756.71
Lance Dunbar5193.80
Phillip Tanner3-1-0.30

Receiving

PlayerRecTgtYdsTD
Jason Witten56670
Dez Bryant46381
DeMarco Murray33280
Dwayne Harris11241
Gavin Escobar12241
Miles Austin25220
Cole Beasley1170

Where this game is likely to land

  • Model
  • Smooth (normal)
  • Betting line

Probability of each final margin. Football scoring clusters — the spikes at 3 and 7 are real, and a smooth model misses them.

line +3.54.1%2.8%7.1%7.5%3.1%4.7%2.9%

Positive margins favour Cowboys; negative favour Rams.

Half-point at risk

2.52%

outcomes that flip if the line moves 0.5

Model home win

53.9%

Expected total

45.9est.

Model

Where the number comes from

Each component is additive, in points of expected margin.

  • Team ratings (opponent-adjusted EPA)-0.70
  • Home-field advantage+2.49
  • Quarterback adjustment0.00
  • Rest differential0.00
  • Expected margin+1.79

One or both ratings are carry-over priors — no games have been played this season yet, so treat this as a starting assumption rather than a read.