Every T20 player, on one scale

Men's T20, 18 competitions, every player with 500 balls or more. Rated in points of win chance against an average IPL player: per innings for a batter, per match for a bowler. Turn the globe, pick a competition, then a player.

How the numbers work

Ratings

A player's rating is what he's worth in win chance against an average IPL player, over the balls he actually gets: per innings for a batter, per match for a bowler. That's why an opener can outrank a finisher who's more dangerous ball for ball: twenty-five balls an innings beats nine.

Behind it, every player has a rate for each kind of ball, from a dot to a six to a wicket, and a measure of how that shifts with the game: the stage of the innings, the rate needed, how set he is, who's bowling. Recent seasons count most; each season back counts about three-quarters as much. For a batter who's still playing, the rating is brought up to his age now: batters hold their level to about 27 and lose a little each year after, so a rating built on his early-thirties seasons comes down to where he is today.

Players who've played in more than one competition are what put all eighteen on the same scale. Not all of a player's record travels, though. Everything is measured in IPL terms, so what he did in the IPL counts in full, and what he did anywhere else counts about three-quarters for a batter and two-thirds for a bowler: that's how much of it carried over when players moved.

Chases
Each chase starts from 300 real IPL moments that match the situation, with the real partner, the bowlers still to come and the venue. From each, the rest of the game is played out ball by ball ten times with him and ten times with an average player: 3,000 games each. Win chance is the share won. The chart on his page draws 1,000 of them.
Ranges
Every rating is recalculated across 120 plausible versions of each player and competition; his range covers the middle 89% of the answers, so a short career gets a wide one. To check them, every player was rated twice, from two random halves of his matches: the gaps between the two came out the size the ranges said they should.
Tested
Rebuilt from earlier seasons only, the ratings were asked to predict each of the last six seasons. They did it better than a player's past win value, his past runs above average or simply his last season, for batters and bowlers alike; once the rating is known, those simpler numbers add almost nothing. A second test rates both real XIs before each match and checks how well that picks the winner; every change to the model has to pass both. In franchise leagues nothing reliably picks the winner better than the toss: not these ratings, not an Elo rating, not last season's record. Auctions, drafts and overseas caps even the sides out too much for any of them.
IPL numbers
Strike rate, economy and the rest are the same player run through today's IPL (the 2026 season's scoring) against average opposition. Tested the same way, they predicted the next season's runs at about the right size.
Next season
A forecast, not a rating. Form fades, so the forecast starts from a version of his rating that leans harder on recent seasons, then scales it back to what typically carries into the next season: about 85% for bowlers, two-thirds for batters. Batters then lose a little for each year they are past 27, on the same curve as the rating.
Fielding
A run-out is credited to the fielders who made it, as win chance per match. Catches aren't: the records don't show dropped catches, so a catch count says more about where a man fields than how well he catches.
Data
Ball-by-ball records from Cricsheet; players with 500 balls or more overall, or 120 in a competition to appear in its list.