Win probability in T20 cricket is not a fixed number — it moves with every delivery, every wicket, every boundary. Understanding how it is calculated helps you read match momentum and spot decisive turning points before the commentators do.

The Starting Point: Pre-Match Probability

Before the toss, CricEdge calculates a base win probability using team ELO ratings, recent form over the last five matches, and venue-specific factors. An ELO gap of 80+ strongly favours the higher-rated team. Gaps under 30 represent genuine toss-ups. Venue history adjusts the probability further: teams that have won 60% or more of their matches at a specific ground get a venue bonus.

Innings One: Setting a Target

During the first innings, the model tracks the batting team's run rate against a projected total based on the venue's average first-innings score. A powerplay score of 60+ in six overs shifts the model toward a high final total. Losing two wickets inside the first three overs reduces projected runs and lowers the batting team's eventual chase-winning probability.

Innings Two: The Chase

The second innings produces the sharpest probability swings. Required run rate crossing 12 per over is a critical threshold — historically fewer than 15% of T20 teams successfully chase targets needing more than 12 per over from the 15th over onwards. The model reflects this with steep probability drops when required rate climbs.

Ball-by-Ball Updates

CricEdge updates win probability after every delivery using its ONNX inference engine. The update latency is under 200 milliseconds, meaning the probability you see during a live match reflects the actual current game state — not a 30-second-delayed approximation from a broadcast feed.