Research · Proposal · Updated Oct 5, 2026

Which bus stops are too close together? About 1,000, on 40 routes.

Closing them would help a little. Close the quieter stops that sit within 750 ft of the next one, on 40 tightly spaced routes in Manhattan, Brooklyn and Staten Island, and riders save about 600 hours a day, even with each minute of extra walking counted twice. Spacing every route to a quarter mile would cost riders about 36,000 hours a day more than it saves.

Every version of the policy, side by side

Net rider-hours each weekday, citywide: time saved on board minus extra walking. Three spacing rules, four stop costs. The cost real redesigns achieved is about 11 seconds a stop; at that cost, closing stops loses riders time under every rule unless the stops closed are the quiet ones.

Walking counts Stops closed are

Where it pays off

The 750 ft rule on the routes where it still pays at 11 seconds a stop. These have the tightest stops, so each rider passes more of the ones that close. Ranked by net rider-hours a day if the stops closed carry half their route’s average riders. The last column says how busy they can be before the route stops gaining.

RouteBoroughStops closedGap, ft Min off a full tripNet rider-hours a dayPays while closed stops carry under

On the M4, every other stop on Broadway

Did it work before?

The MTA has done this twice. The Bronx redesign (June 2022) and the Queens redesign (June and August 2025) both spread stops further apart. Buses got faster, but by far less than the model expected. That is why this page is cautious.

How this works

For every local, limited and Select Bus Service route, we took its usual weekday stop list and measured the distance between stops along the route. Then we closed stops under each rule, and weighed the time saved by riders on board against the extra walk for riders who used the closed stops.

  • The rulesFind the closest pair of stops under the minimum spacing (750, 1,000 or 1,320 ft) and close whichever one leaves the more even gaps. Repeat until no pair is too close. Never closed: the first and last stop, any stop within 100 m of a subway station, and any stop where two or more other routes cross. No closure may leave a gap over twice the minimum.
  • What a stop costsRegressing the MTA’s bus times between timepoints on the number of stops between them gives 30.8 s a stop (95% interval 23.7–37.4). The MTA plans on about 20 s. But the speed gains the Bronx and Queens redesigns actually delivered match only 11 s and 4 s a stop. We use 11 s as the central case.
  • Riders on boardEach route’s weekday boardings (March–August 2026), spread evenly over its stops. Riders passing a stop = boardings × an average trip of 2 miles ÷ route length.
  • WalkingPeople who used a closed stop walk to the nearest one left, at 3 mph. They are spread evenly along the street, so the average extra walk is about half the new gap. Each minute of walking counts as two minutes on the bus, the usual planning weight.
  • The targeted listRoutes where the 750 ft rule still saves time at 11 s a stop as long as the stops closed carry under 60% of the route’s average riders, and that lose at least five stops.

Caveats: there is no public count of riders at each stop, so the model treats every stop on a route as equally busy. The targeted plan only works if the MTA closes the quiet ones, which it can check with its own counts. Walking is measured along the route, not on a street map, and ignores hills, crossings and weather. For older riders and people with disabilities an extra block can be the whole trip: stops by hospitals, senior centres and schools would need protecting, and our rules don’t yet do that. We don’t model waiting: a longer wait in the cold or the dark is a cost, while the bus-hours saved could run more buses and shorten waits. The redesign check compares boroughs, not single routes, and both redesigns changed more than stops.