On June 29, the Better Bus Redesign replaced the entire Metrobus network — every route renamed, most redrawn. DCX is measuring whether it actually made the bus more reliable: how long you'll really wait, on-time rates, and detours, scored from real arrivals — not the schedule, and not WMATA's own numbers. Every Metrorail line too.
The old network's last clean stretch — nine days, June 12–20 — is frozen as the baseline: 125 bus routes measured, 53% on-time, a typical route's expected wait of 18.2 minutes (route median, not ridership-weighted). The new network's clean measurement starts July 28. Because every route was renamed and most were redrawn, the honest comparison is system-level, not route-by-route — and it accrues over months, not weeks. The first clean week, in early August, is an early read, not the final word.
One thing you should know: a schedule-matching fault on our side corrupted bus measurement from June 21 to July 27. Those days are flagged and withheld — dashes, not grades — because publishing them would be fiction. Rail was unaffected. How we measure, including our own failures →
Keyed access to the enriched warehouse — ground-truth arrivals, segment travel times, detour history, and the full prediction benchmark. For operators, researchers, and newsrooms.