Seeking a new way to characterise player skills and styles for viewers, the ICC’s production team commissioned the new Skill Scales for the Women’s T20 World Cup in June 2026. CricViz Head of Data Science Surya Kocherlakota and Broadcast & Product Manager Srinivas Vijaykumar explain the methodology behind the ratings.
The Skill Scales evaluate players across the defining skills and styles of T20 cricket. For batters: Power, Control and Innovation. For bowlers: Threat, Control and Variations. Each is rated on a 0-100 scale, calibrated across all women’s T20 players since the start of 2023 with a minimum of 180 balls faced or bowled.
First, the definitions:
Batting
Bowling
While the skill measures tell us how well a player performs; tendencies describe how she plays. A high Innovation rating doesn’t mean a batter is better, just that she plays more unorthodox shots. A high Variation rating doesn’t necessarily make a bowler more effective; it tells you what kind of bowler she is.
Skill ratings are moderated when they’re based on limited data: for batters with fewer than 1,500 balls faced, and for bowlers who have bowled fewer than three overs per match on average. This prevents small samples from producing extreme ratings.
Why the metrics work this way
T20 cricket is played across leagues, conditions and formats that vary enormously. An economy rate of 7.5 means something very different in matches where a team scores 180 at 9 runs per over than in matches where they score 140 at 7 runs per over. Similarly, an 85% controlled shot rate may be excellent on one pitch and unremarkable on another. Taken at face value, a raw rate can end up telling you more about the circumstances a player found herself in than about the player herself.
The Skill Scales address this by comparing each player against the other players in the same matches she played in, and for bowling, the same phases of those matches. Rather than asking how good a player’s raw numbers were, the ratings ask how much better or worse she performed than her peers under the same conditions. Whatever made conditions difficult or favourable applied equally to everyone there, so comparing within those constraints helps isolate what each player did better or worse than those around her.
The guiding principle in developing these metrics was to start from the simplest possible base rate and add only the adjustment layers that earned their place. Each candidate layer had a principled motivation, and each was validated the same way: our team of editorial analysts reviewed the resulting ratings against their expert judgement of the players, and a layer was kept only where it demonstrably improved that fit.
Six hitting, the basis for Power, needed no adjustment beyond the small-sample correction. Power is closer to a latent attribute than a performance outcome: the raw ability to strike the ball over the boundary. As a result, six-hitting frequency turned out to track that ability well on its own.
Controlled shot rates required adjustment for match context. False shots are heavily condition dependent: facing bowling on a difficult surface produces more of them regardless of the batter’s ability, so without match relative comparison the ratings reflected conditions more than they reflected the batter. With the adjustment in place, they aligned with how our reviewers assess batting from watching the game.
Bowling metrics required adjustment for both match and phase. Scoring rates vary substantially across the phases of a T20 innings, because the quality of the batters at the crease and the risks they are willing to take change as the innings progresses. An over in the powerplay and one in the middle or final overs are simply not comparable on raw numbers, so each bowler is measured against other bowling in the same phases of the same matches. One further layer proved necessary: counting performances against weaker opposition, defined as T20 Internationals involving nations outside the top 10 and lower tier domestic competitions, for 20% less. Without this discount, bowlers who played disproportionately against weaker opposition rated above bowlers who consistently faced stronger sides, a pattern our review process flagged early and the adjustment resolved.
The result is a set of metrics built with the minimum adjustment necessary to make the ratings meaningful.
What the ratings don’t resolve
Match relative comparison handles a lot of the contextual noise that simple rates miss, but it has limits worth being clear about.
The adjustments are calibrated against typical patterns rather than every individual case. A player whose schedule is especially heavy on weaker opposition may still see her ratings flattered despite the 20% discount, and in the same way, the strength or weakness of a player’s teammates can skew the match relative comparisons despite the contextual adjustments. These effects are usually small, but they exist.
Above all, the Skill Scales describe what a player has done across a substantial sample of cricket. They won’t tell you what she’ll do next, which is, after all, why we watch the game.
Leaders in each skill
The top five players in each Skill Scale since the start of 2023, with the underlying rates shown alongside the ratings:
Batting
Bowling
What comes next
The Skill Scales are one part of how we think about evaluating cricketing performance at CricViz, and we’ll have more to share on player valuation in cricket more broadly in future pieces.