● MLB BETTING Q&A · BY MARCDUCK
Why Do MLB Totals Go Over or Under? The Real Drivers
MLB totals don't move by intuition. Park, weather, starting pitcher quality, and bullpen depth drive totals far more than public attention does. Here's what actually moves the over-under needle.
The Top 5 Drivers
- Starting pitcher quality. Two ace SPs (xERA ≤ 3.0) suppress runs by ~30%. Two replacement-level SPs let offenses pile on.
- Park factor. Coors Field plays at 1.23x league average. Petco at 0.96x. Park alone swings totals by 1-2 runs.
- Weather. Each 10°F over 70 = ~0.2 R/team. Wind blowing OUT vs IN at exposed parks (Wrigley) swings totals by 1+ R.
- Bullpen depth. Strong-bullpen teams cap late-inning blowups; weak bullpens let games balloon to 12+ run finals.
- Umpire. Wide-zone HP umpires cap runs by ~0.4 R/game; tight-zone umps add ~0.4. We track 50+ named umpires.
Common Total Patterns
- Day games go UNDER. Visibility cuts runs ~0.6 R/team. Books usually adjust but sometimes lag.
- Coors Field at 11.5 OVERS. Public hammers OVER. Sometimes correct; often line has already moved enough that there's no edge.
- Cold April / October games go UNDER. Cold air = less ball carry = fewer HR.
- Hot August games go OVER. Warm humid air boosts ball flight and stamina-tax bullpens.
- Strong-bullpen teams suppress late OVERS. Brewers, Phillies, Dodgers bullpens cap blowup innings.
Why Most Public OU Bets Lose
Public bets totals on intuition (game "feels high-scoring") or momentum (last game went OVER, expect another). Both are noise. Sharp totals bettors check:
- Park factor adjusted for handedness
- Today's weather vs season norm
- Starting pitcher xFIP gap vs market expectation
- Bullpen quality + recent fatigue
- Day vs night offense effect
- Closing line value — beating the close consistently
How Bookie Bullies Models Totals
Our predicted total = (away_team_neutral_runs + home_team_neutral_runs) × park_factor + weather + wind + umpire ± day/night. Park-neutral team runs strip the team's home park inflation before re-applying tonight's park. We use Negative Binomial distribution (correct family for overdispersed Poisson run totals) instead of Gaussian. Probabilities are blended with the market O/U at 55/45 — the market is sharp but not always right.
When to Bet the Total
Bet OVER when: warm weather + hitter park + below-average SP + hot offense + wind blowing out.
Bet UNDER when: cold weather + pitcher park + ace SP + strong bullpens + day game.
Most days the total has small edge either direction. Lock-tier picks are rare — when 4+ factors all push the same direction AND the model agrees with the market price, we have an edge.
Related