Numbers from the Pitch · Issue 1

How good were the chances? FQ, October 2026

Goalscope · November 2026 · ~3 min read

Why a number like FQ exists

A football score tells you what happened. It rarely tells you how. A team can create five clear chances and lose 1–0; another can have one half-chance and win. Anyone who follows several matches at once knows the feeling: the score says "quiet", the match says otherwise.

Expected goals (xG) was invented to close that gap. It rates every shot by where it was taken from and how, then adds the ratings up. It works well — but it needs shot coordinates, and those arrive late, unevenly, or not at all for many leagues during live play.

FQ — Finishing Quality — is our answer to the same question with a different constraint: it is computed live, from the match statistics that are available while the game is being played, without shot coordinates. It is not a replacement for xG. It is a measure built for the moment you are actually watching.

What we saw in October

We took 20 leagues across Europe, the Americas and Asia, including Türkiye's Süper Lig, and every league match played between 1 and 31 October 2026 that met our data requirements: 651 matches.

74.1%higher-FQ side won (matches not drawn)
80.2%higher-FQ side did not lose
0.57correlation, FQ and goals scored
0.93correlation, FQ and an independent xG model

Three things stood out.

The side with the better chances usually won. In matches that did not end in a draw, the team with the higher FQ won 74.1% of the time. Counting draws as "not losing", the higher-FQ side did not lose in 80.2% of all matches.

FQ tracks goals. Across all 651 matches, the correlation between a team's FQ and the goals it actually scored was 0.57. Football is noisy — deflections, goalkeepers, offside flags — so no chance-quality measure reaches 1. But 0.57 says the number is doing real work.

FQ and xG agree with each other. For the six leagues where we could compare against an independent, publicly available xG model, the correlation between FQ and xG was 0.93. Two methods, two sets of inputs, one of them without shot locations — and they land in almost the same place.

Where FQ falls short

We want to be precise here, because this is the part a newsletter usually skips.

In those same six leagues, the xG model explained goals slightly better than FQ did: a correlation of 0.64 for xG against 0.57 for FQ. That is a real gap, and we do not expect it to disappear. xG knows where each shot was taken from; FQ does not. What FQ offers is not a better answer, but an answer that exists while the match is live, for the leagues we cover, whenever live statistics are available, from the numbers already on the screen.

Put differently: FQ gets within reach of a coordinate-based model using none of its coordinates. That is the result we are proud of, and also the ceiling we are honest about.

What this means when you open the app

When a live match shows one side with a clearly higher FQ, that side has been creating the better chances so far. It is a description of the match up to now — not a forecast of what comes next. Goalscope does not predict scores and shows no odds. FQ is there so that you can read a match in a glance and decide whether it is worth your attention.

Read the full report

Method, confidence intervals, league-by-league tables and the limits of the sample are in the October 2026 FQ report. More issues will follow.

Goalscope — a live football radar, built by statisticians. Free on iOS and Android.