How Sportsbooks Are Using AI to Set Odds in Real Time

From Human Oddsmakers to Machine Models

For decades, sportsbook odds were set primarily by experienced human oddsmakers who adjusted lines based on betting patterns and expert judgment. Today, most major sportsbooks rely heavily on machine learning models that ingest huge volumes of real-time data to price markets automatically.

How It Works

Modern odds-setting systems typically combine:

• Historical performance data across thousands of games
• Real-time in-game data feeds, including player tracking
• Live betting flow to detect sharp money movement
• External signals like weather, injury reports, and lineup changes

These inputs feed into models that continuously update a sportsbook's live odds, often adjusting lines within seconds of a key play.

Why It Matters for Bettors

Faster, more accurate pricing means tighter margins on popular markets, but it also means mispriced lines disappear much faster than in the past. "The days of finding a soft line for very long are basically over on major markets," said one former oddsmaker who now consults for a sportsbook's data science team.

The Human Element Isn't Gone

Despite the shift toward automation, most sportsbooks still keep human traders in the loop to handle unusual situations models haven't seen before, such as an early-season injury to a star player or an unexpected coaching change.

What's Next

As AI models improve, some in the industry expect fully automated trading desks to become the norm for major markets, with human traders shifting toward oversight and risk management rather than day-to-day line-setting.