Cut-n-pasted from a FB page called “The Running Week”
Strava just took a major step to restore credibility to its leaderboards, over 4.45 million activities have now been purged from the platform.
Why such a huge cleanup? A large number of those entries were either misclassified under the wrong sport or captured while the user was actually in a moving vehicle.
The intention behind this move is clear: to ensure KOMs and QOMs reflect genuine human effort, not accidental car rides or mislogged workouts.
Earlier this year, Strava introduced a machine learning system that analyzes 57 different variables including acceleration and speed to detect activity that doesn’t belong. Since launching the system in February, Strava has been methodically re-evaluating leaderboard results to spot any irregularities.
The results speak volumes: millions of dubious entries removed and a significant 72% drop in user-reported “vehicle” activities.
For athletes who’ve battled for segment crowns only to be beaten by an implausibly fast effort, this update is welcome news. Better yet, the AI doesn’t just flag issues — it allows users to correct them by setting an activity to private or deleting it entirely.
But that’s not all. Strava is leaning hard into innovation. The platform now offers enhanced tools for subscribers, including AI-driven route suggestions, interactive POIs like cafés and restrooms, and seamless point-to-point navigation.
And in a signal that they're not slowing down, Strava has also acquired The Breakaway, a cycling training app, further expanding its ecosystem.
Bottom line: Strava is making it clear segment leaderboards should showcase real athletic performances, not shortcuts or GPS mistakes.

Dan