What a win probability curve tells you that a scorecard does not

A cricket scorecard is a precise record of what happened and a poor guide to why. Two batters can score 40 each, one in a lost cause and one when the match hung on every ball. A win probability model tells the two innings apart.

How it works

A win probability model estimates, at any point in a match, the chance that each side will win. It learns from history: thousands of past matches, ball by ball, with the score, wickets, overs and result recorded at every stage.

In the first innings, the model works from runs, wickets and overs, and from where the match is being played. In the second, it adds the most important number of all, the target. Context matters too: the toss and the decision that followed, whether the chasing side is at home, and whether the match is at a World Cup, where pressure changes how teams play.

What the curve shows

Plot the chance of winning ball by ball and the shape of the match appears.

• Turning points show up as steep moves in the curve: a wicket, an over that goes for 20, a dropped catch.
• Pressure shows up as flat stretches where neither side gains, often before a collapse or a surge.
• Contributions can be measured by how much each ball moved the curve. That is a fairer measure of a player's impact than runs or wickets alone.

A scorecard tells you who scored. The curve tells you who won the match.

Why women's cricket deserves its own model

Women's T20 internationals are often analysed with assumptions borrowed from the men's game. That is a mistake. Scoring rates, the value of a wicket, the effect of the toss and the size of home advantage are all different, and they have changed quickly as the women's game has grown. A model trained on women's matches will read a match state very differently from one trained on men's.

The challenge is data. There are fewer women's internationals, and some teams play only a handful of matches a year. That makes it more important to be honest about how confident the model is, especially early in an innings and for less familiar teams.

Why this matters beyond cricket

Win probability is a small, clear example of a broader idea: judging a decision by the situation it was made in, not only by the result. In lending, insurance and investing, the same principle separates luck from skill, and a model that shows its uncertainty from one that hides it.