How Courtside measures market differences
The board separates a raw price gap from a useful signal by accounting for liquidity, freshness, fees, and contract risk.
Build the reference probability
We remove sportsbook margin, weight eligible sources, reject stale or malformed inputs, and calculate a consensus probability. A reference timestamp and source trail travel with every result.
Verify contract equivalence
Venue contracts are matched to canonical NBA outcomes. Overtime, cancellation, push handling, result source, and settlement timing determine whether a match is exact, carries basis risk, or is incompatible.
Measure the visible book
The engine evaluates available ask depth at a standard comparison size, then accounts for venue fees and slippage. Capacity describes visible liquidity; it is not a fill guarantee.
Apply freshness controls
Book and exchange timestamps are evaluated independently. When either side is outside its threshold, Courtside penalizes or suppresses the difference instead of presenting stale precision.
Test whether the difference is useful
A sportsbook consensus is a reference estimate, not a proven win probability. Our prospective research protocol records the first qualifying pregame NBA home-moneyline candidate per game, freezes its inputs, and checks a new quote 10–20 seconds later. It requires three fresh books, exact contracts, visible depth, modeled fees and an execution reserve. Missed entries remain in the record.
Measure outcomes without claiming a winning system
The research workflow compares frozen forecasts with provider-reported final scores, cost-inclusive hypothetical returns, calibration and near-tip references. Paper entries are not real fills; sports results are not confirmation of exchange settlement. Incomplete results and contract changes require review. Courtside has not established a profitable betting advantage, and price persistence alone cannot establish one.