The most common narrative in baseball betting revolves around the starting pitcher. You look at the probable pitchers, see an Ace facing a number-five starter, and instantly assume the Ace's team has a massive advantage.
While starting pitching is the most heavily weighted individual factor in a baseball game, evaluating a matchup strictly on the starters is a trap.
In modern MLB, starters throw fewer innings than ever before. Games are increasingly decided by bullpens.
The shrinking starter
Twenty years ago, a frontline starter was expected to throw seven or eight innings. Today, the league average for a starting pitcher is closer to five innings.
This means that for roughly 40% of the game, the outcome rests in the hands of relief pitchers.
If you back a team solely because they have a dominant starter, but their bullpen is exhausted or ranks in the bottom tier of the league, you are exposing yourself to significant late-inning risk. The starter might leave with a 2-1 lead in the 6th, only for the bullpen to implode in the 8th.
How our model aggregates pitching
Our engine doesn't just look at the ERA of the guy taking the mound in the first inning. We build a composite profile of a team's run-prevention ability.
This includes:
- Season-to-date team ERA and run prevention: This captures the aggregate performance of the entire staff, smoothing out the variance of individual performances.
- Starting Pitcher adjustments: When we run the daily card, we factor in the specific starting matchup. An elite starter provides a definitive bump to a team's baseline strength score for that day.
- Shrinkage: Early in the season, pitching metrics are highly volatile. A reliever giving up 4 runs in his first outing destroys a team's bullpen ERA, but it isn't predictive of the next 150 games. We use mathematical shrinkage in April and May to prevent the model from overreacting to small-sample blowups.
The "Hidden" Bullpen Edge
One of the most consistent ways to find value in the moneyline market is identifying mismatches in bullpen depth.
Casual bettors and public money over-index on starting pitcher name recognition. If Max Scherzer is pitching against a relatively unknown starter, the public will pound Scherzer's team, driving up the price.
But if that unknown starter's team possesses a top-3 bullpen, and Scherzer's team has a heavily taxed, below-average bullpen, the true win probability is much closer than the market implies.
Our model flags these spots frequently. It identifies teams that play complete, 9-inning defense, rather than teams that rely on a starter to be perfect.
When evaluating a pick, remember: the starter only gets them to the 6th. The bullpen has to get the final nine outs.
For informational use only. Past results don't guarantee future performance. Bet responsibly.