Three different times
The scheduled start or off time identifies when an event was due to begin. It does not establish the time a goal was scored or a horse crossed the line. The collection timestamp records when results.fast successfully received the provider record. A provider may also supply its own modification timestamp.
These times answer different questions. A recent collection helps establish that the collection system is running, but it cannot prove that the upstream data is instantaneous or that a racing verdict is official. Read the feed mode and event status alongside it.
Date filters and event-time displays use Europe/London where an unambiguous timestamp is available. Collection times in the status cards use UTC. The difference between those clock readings changes when UK daylight saving changes; it is not itself evidence of a delay.
Why a result can be missing
An empty list can mean the feed is not connected, the competition is outside the account’s coverage, the result has not arrived, or the selected date predates our collected archive. It can also mean that filters exclude the records that are available.
Start by clearing the event search and status filter, then verify the date and meeting or competition. If the selection is correct, check connection health. Do not interpret an empty list as an official statement that no fixtures or races took place.
What happens during an interruption
The collector retains previously stored records when a request fails. This avoids replacing a known result with a blank value simply because of a transport problem. It also keeps the last successful collection time rather than assigning a new successful timestamp to a failed run.
The interface marks the feed stale or unavailable when successful collection is more than ten minutes old or has never occurred. That is a service-health threshold, not a sporting rule or a promise about upstream latency. A historical event can legitimately retain an older collection time while the current feed continues to work.
Later changes and the archive
The same event identifier is updated when a provider returns a corrected record. Saved links therefore point to the current stored version, not a snapshot frozen at the moment of saving. An early order and a later fuller result should not become separate races.
The racing API collector requests today’s results, while research looks back 48 hours; the football collector requests a short window around the current day. The system does not currently promise to capture all later amendments to older events. If you need to establish a final official record, check the relevant authority and prepare a correction report for any discrepancy.
Provider information
Football collection reads six ESPN public league scoreboards every minute and retrieves available match summaries for games in progress and completed games. Match pages check for stored updates every 15 seconds. Scores are collected snapshots, not an instantaneous service. Score stages and shoot-out values stay unavailable when the source does not provide them. Unavailable league requests are reported as partial coverage.
Racing API records are supplied through The Racing API where account coverage permits. An hourly Codex research job searches official authorities and independent topical sources across the four racing regions, then runs a separate verification pass before storing factual results. Web-researched records are labelled and cannot establish comprehensive or live coverage. The free results integration provides basic fields; a missing margin or condition stays unavailable rather than being inferred. Provider account rights and any additional attribution requirements must be verified before public activation.
For practical examples, read why a score can arrive late, or use the glossary for a shorter definition.