Definition
What is Bayesian pricing in sports betting?
Last updated
Definition
Bayesian pricing computes a fair probability for an outcome by starting from a prior belief (e.g., 'this team wins X% of the time at home against this opponent strength') and updating it with new evidence (current form, injuries, weather) using Bayes' rule.
Sports outcomes have natural priors — historical base rates, ELO-style strength estimates — and incremental information (team news, lineup, weather) updates those priors. A Bayesian pricing model is explicit about the prior and the update; a frequentist model often hides the prior. Glitch Edge's cricket + NBA models use Bayesian-flavored updates: a prior on team / lineup strength gets updated as match state and game-day data arrive.
Why Bayesian fits sports
Most decisions in sports betting are “what should I update on this new info?” Bayesian framing makes the update mechanical: prior × likelihood ∝ posterior. The discipline forces you to write down what you believed before the news arrived, which is the most common modeling mistake retail bettors make.
How Glitch Edge uses it
Cricket: prior on team strength (Elo-style) × likelihood of current state given strength → posterior win probability. NBA: prior on lineup strength × likelihood of current pace given lineup → posterior pace + spread.