Skip to main content

How accurate are the USA - MLB odds?

The closing prices of 181 finished games since Fri 19 Sep 2025, set against what happened: the chance the market gave, with the bookmakers' margin removed, next to how often it came in. A record of what the market did, not a tip.

The market's record

181 games, 1 bookmakers
  • Closing favourite won 59% 106 of 181 games
  • Market Brier score 0.476 0 is perfect; a coin flip scores 0.500
  • Average closing payout 95.9% 100% minus the bookmakers' margin
  • Games covered 181 Fri 19 Sep 2025 to Thu 29 Oct 2026

1,819 of the 2,000 games finished in that time are left out: no bookmaker had a closing price on every outcome.

Market chance against what happened

30-40% chance: happened 31.7% of 41 selections 40-50% chance: happened 46.7% of 120 selections 50-60% chance: happened 52.8% of 142 selections 60-70% chance: happened 68.3% of 41 selections 100% 0%
Market said 0%100%
Happened
Market chancePriceSelectionsMarket saidHappenedDiff
30-40% 2.50 to 3.33 41 36.1% 31.7% -4.4 pts
40-50% 2.00 to 2.50 120 45.3% 46.7% +1.4 pts
50-60% 1.67 to 2.00 142 54.0% 52.8% -1.2 pts
60-70% 1.43 to 1.67 41 63.9% 68.3% +4.4 pts

Every priced outcome of every game is one selection. "Market said" is the bookmakers' average implied probability with the margin removed; "Happened" is how often that outcome came in. On the diagonal the market was right on average; above it, that kind of result happened more often than the prices said. Bands with fewer than 30 selections are left out.

Home, away

OutcomeMarket saidHappenedDiff
Home win 54.0% 56.4% +2.4 pts
Away win 46.0% 43.6% -2.4 pts

The market's average closing chance for each outcome over all 181 games against the share of games it won. "Home" is the first-named side.

Bookmaker accuracy

lower is better
BookmakerGamesBrier scorevs market averagePayout
Bet365 181 0.476 0.000 95.9%

Each bookmaker's own closing prices, margin removed, scored against the results of the games it priced; "vs market average" compares it with the bookmakers' average on the same games, so a negative figure means its prices were the closer. Bookmakers with fewer than 50 priced games are left out. Differences of a few thousandths are noise; a much higher score usually means the bookmaker's last price was captured well before the start, so it is the price in our archive being scored, not the trader.

Odds accuracy FAQ

How accurate are USA MLB odds?
Across 181 finished USA MLB games since 2025, the closing favourite won 59% of the time. The market's closing chance, with the bookmakers' margin removed, gave the home side 54.0% on average and home wins happened 56.4% of the time. The bands above show the same comparison price by price: a well-calibrated market has outcomes happening about as often as its prices say.
What is a Brier score?
The squared distance between the chances a bookmaker's prices gave and what happened, averaged over games: 0 would be perfect foresight, and on a three-way market a coin-flip price scores 0.667. The market average's closing prices score 0.476 in this league; the lowest single bookmaker is Bet365 at 0.476. Differences of a few thousandths between bookmakers are noise, not skill.
Does a well-calibrated market mean the odds cannot be beaten?
Calibration says the market is right on average, not on every game. A band where outcomes happened more often than the prices said is where the closing market under-priced that kind of result over these seasons. The bookmaker's margin, about 4.1% here, still has to be cleared: the closing line value calculator shows what a price needs to beat.