The Control Index asks one question: was this side running the game, or chasing it?

The top of the ratio — big chances created, shots on target — is what a team accumulates when it controls play. The bottom — fouls and cards — is what it accumulates when it doesn’t, and has to stop the game to cope. Dividing one by the other gives a number comparable across fixtures, because it is a ratio not a count.

Why it leaves the score out. The Power grid and Power score above are built from big chances, goals scored and goals conceded — who tried, who scored, who kept them out. That describes a season well, but two of those three are the scoreline, so they largely restate the table. This index deliberately ignores the score, which is what lets it look ahead: across the 2025/26 Premier League, a club’s Control Index over its first 19 games predicted its points in the last 19 better than anything else we tested (ρ = 0.62) — better than the table itself at the halfway point (0.43), and better than the Power score (0.48).

Why chances and not possession. Within a single match, possession has no relationship with the result (r = −0.01) — a side that goes ahead often hands the ball back. Over a whole season it is a reasonable marker of a good team (0.59), but because it is counted in tens it swamps whatever sits beside it. Built on chance quality instead, this index tracks a single match’s result at r = 0.37.

Reading it. A high figure means territory and chances without needing to foul. A low one means ceding the ball and interrupting play. The line is a 5-game rolling average, so it shows the trend rather than the noise of a single afternoon.

Early in a season, read it gently. Fouls sit in the denominator and vary a lot between clubs, so over a handful of games a disciplined side can rank high and a combative one low on discipline as much as control. It settles as the sample grows.

It does not decide results. A side can control a match at 12.0 and lose to a counter-attack, or sit at 3.0 away from home and win. It describes the character of a performance — which is what makes it useful for spotting when a run of results is flattering a team, or hiding how well they are actually playing.

Every input is a core match statistic, so it works for any club in any competition — no expected-goals model required.