Counting Rules and Comparability: a data reference

Two numbers are comparable only if they were counted the same way. This page sets out the rules this site uses, why they were chosen, and what a reader should check before comparing any figure here with a figure from somewhere else.

The general principle

A figure is comparable with another figure when both were produced under the same definition. Where the definitions differ, the comparison is between definitions rather than between the things described, and the result is meaningless however carefully it is computed.

That principle sounds obvious and is violated constantly, because captions are ignored and figures travel. A career total quoted in one article and compared with a career total quoted in another is a comparison of two rules until both are read, and the disagreement that follows is an artefact rather than a finding.

The metric rules

For team and player metrics this site uses a single provider family per table, states the provider family in the caption, counts only senior competitive matches, and reports rates per ninety minutes alongside every total. Penalties are reported as a separate line rather than folded into open-play figures.

Those choices are not the only defensible ones. They are chosen because they make the metric maximally comparable with the widest range of published work, and because each of them is stated on the page where it applies rather than buried in a method note.

The record rules

For record material this site counts senior competitive matches for club and country from the first professional season, excludes youth, reserve, exhibition and tour matches, notes the treatment of wartime and regional competitions explicitly, and reports a range where the classification is genuinely disputed.

Records are checked once a year, active-career figures are labelled as readings at the check date, and no figure is described as official. The reasoning behind each of those choices is set out on the pages that describe the record classes.

The uncertainty rules

Every figure that rests on a small sample is marked as such. Where a rate is computed from fewer than a threshold number of events, the caption says so, and the figure is presented as descriptive rather than as evidence about ability.

The same applies to kept-team and goalkeeper metrics, where the event counts are naturally small, and to any comparison across competitions, where the underlying definitions may differ. Marking the uncertainty is not hedging; it is a statement about what the number can bear.

What a reader should check

Before comparing two figures, check three things: the counting rule, the competition class and the date of the reading. Those three resolve the overwhelming majority of apparent contradictions, and where they do not resolve a difference, the difference is either definitional in a way the captions do not state or it is small enough to be within rounding.

Where a figure arrives without a caption, the safe treatment is to use it only internally: within its own table, for ranking, without comparison to anything external. That is a real limitation and stating it is more useful than pretending the figure is portable.

Why the rules are published

A reference that publishes its rules can be disagreed with precisely. A reader who prefers a different rule can recompute the table from the stated method and reach a different answer that is also defensible, which is a productive disagreement. A reference that publishes only numbers generates arguments that cannot be resolved.

That is the whole justification for the method pages on this site. The rules are not a legal formality attached to the tables; they are the part of the content that makes the tables usable by anyone other than the person who assembled them.

Key reference points

• Two numbers are comparable only if counted under the same definition.
• Metrics use one provider family per table, senior competitive matches and rates per ninety.
• Penalties are reported separately from open-play figures everywhere.
• Records use one stated rule with ranges where the classification is disputed.
• Small samples are marked as descriptive rather than as evidence.
• Check the rule, the competition class and the reading date before comparing anything.

A published rule is what turns a table into a citable reference, because it allows a reader who disagrees to reach his own answer instead of arguing about an unknown.