You check a live match, note that possession reads one way, glance back a few minutes later, and the same game now shows different numbers. It is a common source of confusion for anyone following football in real time. Live statistics do not sit still, and on RubiScore as on any serious data platform, that movement is deliberate. Understanding why the numbers change is the key to reading them well.
Fans notice it constantly. A team is shown with a commanding share of possession in the opening exchanges, and by half-time that figure has drifted toward the middle. Expected goals ticks upward through the game and never comes back down. A pass-completion percentage that looked elite early settles into something more ordinary. The instinct is to assume an error somewhere — that one of the two readings must be wrong.
Almost always, neither is. The changing numbers are not a sign of a faulty feed. They are the natural behaviour of data describing an event that is still happening. A live statistic is a running estimate, and an estimate of an unfinished match is supposed to move.
The single most useful idea here is that in-play numbers are provisional by definition. During a match, every figure is a snapshot of a game still in progress. It is accurate for the moment it describes and provisional about everything still to come. The possession reading at the hour mark is not a draft of the final figure; it is a true description of the match up to that minute. When it changes, it is not being corrected — it is being extended.
That reframing matters because it changes what you should expect from a live page. The number is not trying to predict where it will end up. It is reporting where things stand right now, and "right now" is a moving target for ninety-plus minutes.
Possession is the clearest example. It is a cumulative ratio: the share of time or passes one team has held across the match so far. Early on, that ratio rests on very little play, so a short spell of keep-ball can push it to a lopsided figure. As the minutes accumulate, each new passage of play matters less to the running total, and the number steadies.
So a side shown with a huge early share of the ball, drifting toward parity by full-time, has not seen its data corrected. Both readings are honest snapshots of different amounts of football. The early figure was volatile because it was built on a small slice of the game; the later one is calmer because it rests on the whole thing. Reading an opening-quarter percentage as though it were settled is one of the easiest mistakes to make.
Expected goals behaves differently, and its behaviour trips people up in the opposite way. Expected goals accrues. Each shot adds its estimated value to the running total, and nothing ever subtracts from it, so within a single match the figure only rises. A quiet game can sit almost still for long stretches, then jump when a clear chance is taken or spurned.
This is why a late flurry can rewrite the expected-goals story of a match in minutes, and why the figure at any moment reflects only the chances created up to that point. It is not forecasting the rest of the game. On a platform such as RubiScore, watching that tally grow is a way of tracking how a match is actually developing — but only if you read it as a cumulative record rather than a verdict delivered early.
The other moment numbers move is after the final whistle, and this is where the distinction between provisional and settled data becomes important. Live figures are logged at speed while the game is unfolding. Once it ends, a match record is reviewed and finalized, and small adjustments are normal:
None of these revisions signals a weak feed. The opposite is true: reconciliation is a quality-control step, the point at which fast live data is checked against a settled record. RubiScore treats the provisional live figure and the finalized post-match figure as two stages of the same process, not as a mistake followed by a fix.
Underneath all of this sits a single tension that every live data service manages: speed against precision. A live feed prioritizes latency, because a statistic that arrives ten minutes after the event is useless to someone watching in real time. A settled record prioritizes accuracy, because there is now time to review every decision carefully. You cannot fully maximize both at the same instant, so the sensible design is to serve fast provisional numbers during play and reconcile them into precise ones afterward.
That is the approach the platform takes. During the match, the RubiScore feed favours immediacy so the page keeps pace with the game. After it, the same match is settled into a finalized record. The movement you notice between the two is the visible edge of that trade-off, not a flaw in it. A service that never revised anything would either be slow in play or wrong afterward.
Once the behaviour makes sense, a few habits turn it to your advantage:
Follow those and a live page stops being confusing and starts being genuinely useful. You are no longer waiting for the numbers to hold still; you are reading them for what they are at each stage.
Not everyone reads a live page for the same reason, and the provisional-versus-settled distinction matters more to some followers than others.
For anyone tracking a game in play, immediacy is everything. A supporter watching at the ground or on a stream wants the numbers to move with the match, and the occasional wobble in an early percentage is a fair price for staying level with events as they happen. For this reader, the provisional feed is exactly the right tool, as long as they treat its early figures as the moving estimates they are.
Fantasy football managers sit at the other end. Much of what they care about — an assist credited to the right player, an appearance that counts, the bonus-style adjudications some competitions apply — is only finalized after review. Reading those off a live feed and reacting instantly invites frustration, because the provisional version may still change. The settled record is the one to act on.
Analysts and anyone comparing teams over time need the finalized data most of all. Comparisons only hold when every match in the sample has been settled to the same standard; mixing a mid-game snapshot of one fixture with the completed record of another produces nonsense. For serious comparison, the provisional feed is a live convenience and the settled record is the source of truth. The platform serves both needs on purpose, which is why the same match can look slightly different in the ninetieth minute than it does the next morning.
Live football statistics change because they describe a live event, and then they settle because a finished match can finally be reviewed in full. Both stages are correct for their purpose: the provisional figure keeps you level with the game, and the settled figure gives you a record you can trust and compare. Seen that way, the movement is a feature rather than a bug — the natural rhythm of data that has to be both fast and, eventually, exact.
RubiScore is built around that rhythm, streaming provisional numbers live and reconciling them into a finalized record once the whistle goes. Both the in-play and the settled figures for matches, teams and players are published on rubiscore.com.


