DAL @ HOU
Week 4 · Sun 1p · All games · Talk about this game
HOU −2.5
spread
47.5
total
25 – 22.5
score implied by the line
The sportsbook's line, not our projection: (total ± spread) ÷ 2 gives each team's share of the points. The ladders below start from these lines; our own number comes next.
Our projected spread
HOU by 6.0
ours
HOU by 2.5
the line
We're 3.5 points more on HOU than the line. The biggest reason is the power ratings (HOU #11 on the field but #9 in the market; DAL #25 vs #18). The market prices things our formula doesn't see: injuries, matchups, weather, last season, and the sharpest money in the world.
Is a gap this big an opportunity? When we've disagreed with the line by 3–6 points, our side covered 44% of the time (399 games, 2019–2026). You need 52.4% to break even at −110. The line has been the better bet.
For your bet: This number explains the game; it isn't a bet against the line. Tested on 319 games it never saw (2025–2026), the line was closer to the final margin 60% of the time, and betting our number against it won only 43%. The full record
Spread ladder
For your bet: Dog +3.5 wins 59%, +2.5 only 50% (9% land on 3). The hook is worth paying for up to −141.
Show the 1,302 games behind these numbers
Every NFL regular-season game 2006–2026 where the favorite laid 2 to 3 points (as close to 2.5 as gives at least 300 games), and how it finished. Each bar is how many games the favorite won by that margin; below zero, the underdog won.
- Chance: the share of those games. “HOU by 7+” = games won by 7 or more ÷ all 1,302.
- Fair odds: the price with no bookmaker cut, where you'd break even over time. 50% is +100; 25% is +300; 75% is −300. Pushes are set aside.
- Key numbers (outlined): 3 and 7 are the most common margins, a field goal and a touchdown.
- Checked: we re-test this on every update. See how close it's been.
Your line
50%
win
0%
push
50%
lose
Fair price +101 · at −110 it's overpriced: about −5¢ per $1
Pick the bet your book offers and type its price. Win, push and lose are the share of the same past games. Value is the chance of winning × what you'd win, minus the chance of losing × your stake; within 2¢ per $1 we call it fair.
Total ladder
For your bet: Totals have no big key numbers: no single final total happens much more than 4% of the time. Don't pay extra juice to move a total by half a point.
Show the 2,690 games behind these numbers
Every NFL regular-season game 2006–2026 with a half-point total, and how far the final score landed from its posted total. Shown here as if the total had been 47.5. Totals move with the era, so the line itself sets the middle, and history sets the spread around it.
Touchdowns move this game
If he scores…
- Javonte Williams scores → DAL covers 60% (vs 42% when their RB doesn't)
- David Montgomery scores → HOU covers 60% (vs 42% when their RB doesn't)
- Dak Prescott throws 2+ TDs → the over hits 68% (vs 34%)
- C.J. Stroud throws 2+ TDs → the over hits 68% (vs 34%)
These are averages for every starting RB (3,277 games) and QB (3,631 games) 2019–2026, not numbers for these players. The players are this week's likeliest scorers on Boom TD!.
For your bet: These legs move together. An anytime TD plus his team covering hits about 19% more often than the two chances multiplied, so it's a natural same-game parlay. Check the parlay price against that.
Power rankings
Ranks explain the line; the line is the better predictor. How we rank · All 32 teams · This week's edges