How every figure on this site gets here, and what we refuse to do.
Everything on a race page is one of three things, and they never get mixed. Official facts come from the race's own published sources and carry a link to them. Derived figures are arithmetic over official facts, and we say which facts. Participant reports are what people who were there say happened, labelled as observation and never promoted into race policy.
When a race does not publish something, the row says so, and where we looked. It does not say "none", and it does not quietly become a reassuring blank. There is a real difference between a race that states it has no intermediate cutoffs and a race that has simply never mentioned them, and that difference can decide whether you get pulled at mile 20.
We do not publish a walker-friendliness score. Any single number would be a judgment dressed up as a measurement, and we have neither the participant data nor the standing to make it. We do not estimate your pace from another race. We do not use the word safe. And we do not compute a figure for a mechanism we have not verified — a race with a rolling sweep gets its facts displayed and no calculator, because a confident wrong answer is worse than none.
Races close their courses in more than one way. A fixed time limit is a duration measured from the gun or from your own start. A rolling sweep is a vehicle moving at a set pace, anchored to something — often the last starter, which matters enormously. Large races frequently run both at once, and the two can give different answers for the same participant. Where both exist we show both and never collapse them into a single verdict.
The number is not the point. What happens if you miss it is the point, and races say several different things: you may be removed from the course, you may finish without an official time, you may finish unsupported. Where a race hedges — and they usually do — we keep the hedge instead of flattening it into a rule.
Every record shows the date we last checked it. Records older than 270 days drop out of the finder's default results and carry a warning, because a stale cutoff is worse than no cutoff. The database is small deliberately: we would rather list 2 races we can keep current than several hundred we cannot.
If you are a race organiser and something here is wrong or out of date, tell us and we will publish your statement alongside the record, dated. We will not delete participant reports that disagree with it — we will show both and say they conflict.