When a Finished Match Changes: RubiScore's Approach to Post-Match Statistical Corrections

A completed football match is not always a settled one. In the hours and days after full time, a goal can be reassigned to a different scorer, an assist can appear or vanish, and a red card can be removed from a player's record entirely. The scoreline almost never moves. Almost everything attached to it can.

This is a different phenomenon from the one most readers have noticed. Numbers moving during a match are the visible kind of change, and they stop at the final whistle. The corrections described here begin after it, and they are far less visible, because nobody is watching the page when they land.

Why a finished match is still provisional

The record of a match is not produced by a single authority. The referee files a match report. The competition organiser publishes an official set of statistics. A disciplinary body may later rule on incidents the officials flagged or missed. Data providers collect and structure all of it. Each of these bodies works to its own timetable, and the record only becomes final when the slowest of them has finished.

RubiScore, like any provider working from official competition sources, therefore publishes two things that look identical but are not: a full-time record, which is complete for the purposes of watching football, and a settled record, which is what the competition will still be asserting a week later. The gap between them is usually small. It is not always trivial.

What actually gets amended

Four categories account for nearly all post-match changes, and they are worth separating because they behave differently.

• Goal attribution. A shot deflected on its way in raises a genuine question of whose goal it was, and different competitions resolve it differently. Some allocate it on the spot and leave it; some review contested cases formally, with a panel that can reassign a goal to the deflecting defender as an own goal or back to the original shooter days later.
• Assist attribution. Assists are a convention rather than a Law, which makes them the most amendable field in football data. Whether a deflected pass, a defensive error between pass and shot, or a shot rebounding to a scorer counts as an assist is a definitional judgement, and providers and competitions do not always reach the same one.
• Disciplinary records. A red card overturned on appeal is removed from the player's suspension and, depending on the competition's convention, may or may not be removed from the match record. Retrospective action runs the other way: a card that was never shown can be added to a player's disciplinary total after a review, without ever having existed on the pitch.
• Mistaken identity. Rare, but the cleanest case of all — a card or a goal recorded against the wrong player, corrected once video makes it obvious.

Changes to the result itself belong to a separate category, governed by competition regulations on abandonment, ineligible players and awarded matches, and they are far rarer than any of the above.

Who has the authority to change a number

This distinction is the one most often missed. A data provider does not adjudicate a contested goal. It mirrors the decision of whichever body owns the official record for that competition, and it updates when that body updates.

The practical consequence is that identical incidents can be resolved differently in different leagues, because the deciding body is different in each one. A competition with a formal review process will produce a reassignment that a competition without one will never make. Neither is wrong; they are running different conventions, and the difference belongs to the competition, not to the data.

It also means that a provider disagreeing with an official source is a bug rather than an opinion. The RubiScore platform's job is to reflect the record accurately and to reflect changes to it promptly, not to hold an independent view about who scored. Where a competition publishes no formal review mechanism at all, the original allocation simply stands, and the honest thing for a data platform to do is carry it rather than quietly impose a house rule that would make the same incident read differently from one league to the next.

How long the window stays open

There is no universal deadline, but the timings cluster.

• Most attribution corrections land within seventy-two hours, while the match is still being reviewed and before the next round of fixtures.
• Disciplinary appeals follow the competition's own procedural calendar, typically within a few days, because a suspension has to be resolved before the next match.
• Formal goal reviews, where they exist, may be batched and can take longer, occasionally being confirmed weeks later.
• A small number of amendments are only made at the end of a season, when a competition reconciles its official records.

That last category is the reason a historical season table and a contemporaneous match report can disagree without either being careless. It is also why Rubi Score records for a concluded season are a safer citation than a figure captured on the day, and why serious analysis of individual output is usually run once the competition has closed its books.

What this does to a data series

The consequences for anyone using football statistics are larger than the size of the corrections would suggest.

A player's season totals move silently. If a goal is reassigned two days after a match, every derived number changes with it — goals, goals per ninety, conversion rate, and any ranking the player appears in. Nothing announces this. A total quoted on Sunday and the same total quoted on Wednesday can differ with no error at either end.

Screenshots age badly. A figure captured at full time is a legitimate record of what was published at full time, which is not the same as a record of what happened. Analysis that depends on precise attribution should be run after the amendment window, not during it.

Sources disagree for timing reasons more often than for quality reasons. When two platforms show different assist totals, the usual explanation is that one has already ingested an amendment and the other has not, or that the two are working from different official sources with different conventions. Treating that as evidence of inaccuracy is usually wrong.

Aggregate metrics absorb it better than individual ones. A team's expected goals or shot count is essentially unaffected by whose name is attached to a goal. Individual scoring and creative records are directly exposed. The rule of thumb is that the more granular the credit, the more vulnerable the number.

How the platform handles the gap

RubiScore's approach is to treat the official competition record as the source of truth and to carry corrections through rather than freeze a match once it ends. Match pages, player season totals and competition-level records are all recomputed when an upstream amendment arrives, which is what keeps a player's aggregate consistent with the sum of the individual matches behind it.

The honest limitation is that no provider can move faster than the body making the decision. If a competition confirms a reassignment on the Thursday, the record before that Thursday was not inaccurate — it was accurate to the best official information then available. Anyone building on this data should understand that a completed match is a claim with a timestamp rather than a permanent fact.

What to check before quoting a statistic

• Confirm the match is outside the amendment window if the number depends on individual attribution rather than team totals.
• Note the competition, because review conventions vary and a goal treated one way in one league may be treated differently in another.
• Prefer the current published record over a saved screenshot or an old report when the two conflict.
• For disciplinary counts, check whether an appeal was heard, since a rescinded card may sit differently in the match record and the suspension record.
• For research spanning a full season, pull the data once after the season closes rather than accumulating it week by week.

The misconceptions worth clearing up

The first is that corrections are evidence of sloppiness. They are the opposite: a record that never changes is a record that is not being reconciled against anything.

The second is that assists are officially defined in the same sense that a goal is. They are not. There is no Law of the Game covering assists, which is precisely why they are amended more often than any other attacking statistic and why two respectable sources can hold different figures indefinitely.

The third is that this is a modern problem created by video review. Contested attributions long predate it. What has changed is the speed of detection and the visibility of the correction, which makes an old process feel new. Match records, player totals and competition histories, kept in line with official amendments rather than frozen at full time, are published on rubiscore.com.