Prompt
Why do my routing ETAs look wrong in certain cities?
Latest observation
Oct 15, 2026
OpenAI APIWeb search: off
Routing ETAs can look “wrong” in certain cities for a few common reasons, especially in dense urban areas:
- Traffic is highly localized and fast-changing: Downtown grids, event traffic, school zones, and rush-hour spikes can make average travel times unreliable.
- Map data may be incomplete or stale: New roads, one-way changes, closures, construction, or lane restrictions can throw off routing.
- Turn delays matter more in cities: Left turns, traffic lights, stop signs, and intersections add significant time that a simple distance-based estimate may miss.
- Restricted roads and access rules: Bus lanes, tolls, delivery windows, private roads, and pedestrianized streets can affect routing if not modeled correctly.
- GPS/map matching issues: In areas with tall buildings (“urban canyon”), location fixes can drift, causing the route to snap to the wrong street.
- Pickup/dropoff and parking time: The ETA may reflect driving time only, not the extra time to find parking, enter a building, or navigate complex curb access.
- Historical data bias: If the model is trained on broader regional averages, it may not capture city-specific patterns well.
If you want, I can help you diagnose it more specifically. A few useful details would be:
- the city or cities,
- whether the issue is with driving, delivery, or fleet routing ETAs,
- whether ETAs are consistently off by a fixed amount or only in certain conditions,
- what routing engine or data source you’re using.