How long you actually wait

The gap between trains, measured rather than timetabled.

Embed this viewCheck freshness

Data access

Use the machine-readable JSON record or the flattened CSV download or the compact Markdown version.

Methodology

Median gap between consecutive arrivals per line, computed from live predictions across every station reporting at least four of them. Gaps over twenty minutes are discarded as the end of a timetable window rather than a service interval. A line can run plenty of trains and still make you wait if they are bunched, which is what this measures and a train count does not.

Provenance

Source
TfL API
Cadence
live
Engine
live-feed

Keep going