Your top five
About these records and the replay
We search the original watch cumulative-distance stream for the fastest continuous 5,000m, 10,000m or 21,097.5m in each activity. Boundary times are interpolated between samples. Stopped time inside a segment remains included; GPS is not scaled to the watch distance and average pace is not used to invent a personal record or animate a runner.
The five ghosts use their native distance/time samples, clipped to each calculated segment. The course is a shared virtual illustration, not a shared GPS route. A ghost can lead early and still finish slower. The timing tower stays ranked by finish time; a separate leader badge follows the clock. The row highlight only identifies your selection.
Calculated rolling times are marked ≈. Separate activities only; overlapping windows from the same run do not fill the rankings. These are watch-distance estimates, not certified race results or a reproduction of Zepp’s proprietary records algorithm. Native recorded laps appear separately in the source-run panel. Each effort must fit entirely inside one verified stretch. Gaps and faulty sections are excluded without discarding intact segments elsewhere. Unverified pause clocks or faults that cannot be located may still prevent a record.
The selected ghost's red number is its recorded heart rate at that replay point, aligned to the segment's source timestamp. Gaps over ten seconds, invalid readings and unverified sensor clocks show —. The heart pulse is decorative. Source-run heart rate and cadence below remain whole-run averages. Weather is an hourly model estimate averaged across the full run. Route, conditions and effort vary between dates; a time comparison alone is not a fitness assessment.