MLB betting angles — tested out of sample
13 pre-defined angles were graded at closing prices on 2016 through 2021 (in-sample) and then, unchanged, on 2022 to date (out-of-sample). An angle qualifies only if it shows a profit at the actual closing odds in both periods with at least 300 / 100 bets. Most don't — that's the point. 0 of 13 qualified.
No upcoming games match a qualifying angle right now.
Every angle tested
| Angle | Market | In-sample W-L-P | Win% | ROI | Out-of-sample W-L-P | Win% | ROI | Qualifies |
|---|---|---|---|---|---|---|---|---|
| Postseason underdogs | ml | 65-110-0 | 37.1% | -14.4% | 82-93-0 | 46.9% | +5.4% | no |
| Underdogs on a 4+ game losing streak | ml | 0-0-0 | 0.0% | +0.0% | 0-0-0 | 0.0% | +0.0% | no |
| Moneyline favorites −200 or shorter | ml | 1514-619-0 | 71.0% | +0.3% | 1329-588-0 | 69.3% | -2.4% | no |
| Underdogs after a loss by 5+ runs | ml | 838-1256-0 | 40.0% | -1.3% | 721-1040-0 | 40.9% | -2.6% | no |
| Unders, totals of 10 or higher | under | 842-800-65 | 51.3% | -1.9% | 418-401-27 | 51.0% | -2.7% | no |
| Run line underdogs (+1.5) | ats | 4639-3656-0 | 55.9% | -2.2% | 2270-1812-0 | 55.6% | -3.4% | no |
| Fade teams that won by 5+ runs last game (bet opponent ML) | ml | 1771-1968-0 | 47.4% | -2.2% | 1664-1791-0 | 48.2% | -3.4% | no |
| Home moneyline underdogs | ml | 1678-2426-0 | 40.9% | -5.4% | 1500-2089-0 | 41.8% | -4.4% | no |
| Road moneyline favorites | ml | 3034-2282-1 | 57.1% | -1.4% | 2940-2393-0 | 55.1% | -4.6% | no |
| Overs, totals of 7.5 or lower | over | 1250-1171-66 | 51.6% | -1.3% | 1827-1822-77 | 50.1% | -4.6% | no |
| Home run line favorites (−1.5) | ats | 2203-2936-0 | 42.9% | -6.3% | 1053-1354-0 | 43.7% | -4.7% | no |
| Moneyline underdogs +150 or longer | ml | 1461-2874-0 | 33.7% | -5.3% | 1032-2070-0 | 33.3% | -6.6% | no |
| Unders, postseason games | under | 99-94-7 | 51.3% | -2.2% | 99-104-5 | 48.8% | -7.2% | no |
Odds come from ESPN-listed closing lines and prices (spreads/totals without a listed price are graded at −110; run/puck lines without a price are excluded). Where a game had no posted price it is excluded. Passing both halves lowers the odds that an angle is pure noise; it does not prove an edge, and books adjust.