I was wondering whether it would be faster to walk with or against lights in New York so I made a little simulation to see. Turns out it matters a lot – Around a 30-40% difference at worst case scenarios, but depending on small implementation details going with/against can be super good or super bad.

Assumptions that are probably incorrect

  1. Symmetric Lights – that is a light is green as long as it is red. Generally Avenue Lights are green longer than red
  2. No jay-walking. We can also account for this by extending the amount of time for green vs red.

Using the data from 5th ave and 45th we have a 45 second light with 5 second delay which would mean that you loose 11% of time by walking with traffic. durations_speed_3.1 light_seconds_speed_3.1

The first graphs were assuming walking at 3.1 MPH (19 minute miles). If you raise that to 4.0 MPH (15 minute miles) All of a sudden you hit the sweet spot at 45 second lights and get huge gains. I don’t think you can say categorically it is always better because small differences in any of the variables make a big change in efficiency. Here are the graphs for 4.0 mph.

durations_speed_4.0.1 light_seconds_speed_4.0.1

But a good rule of thumb is fast walkers walk with traffic, and slow walkers walk against traffic. speed_vs_time