February 1


4-Way Stop vs. Roundabout – MythBusters Experiment

By Mike Spack

February 1, 2017

4 Way Stop Sign, Mythbusters, Roundabout

MythBusters set up an experiment to determine which intersection works better – 4 Way Stop Signs or Roundabouts – check out the ten minute video below.

Experiment Setup

  • In a parking lot with barrels
  • Single lane approaches at four legged intersection
  • Show driver which way to go (left, through, right) as they pull up to the intersection
  • Test for two 15 minute periods and present average
  • Test all way stop, uncontrolled with “traffic cop”, and roundabout (let drivers warm up in roundabout for a half an hour since we’re Americans, arguably not used to roundabouts)
  • Looked like about forty cars were circulating, enough to keep queues on all four legs
  • Count the total number of cars that goes through the intersection during 15 minute period


Roundabouts win with 460 vehicles going through the intersection during the 15 minutes.  All way stop signs had 385 vehicles go through in the 15 minutes and the traffic cop had 289 vehicles go through.  Roundabouts also have a safety advantage with less severe crashes.

Of course stop signs are a lot cheaper than a roundabout, so there are plenty of intersections that don’t warrant the extra cost of a roundabout but are just busy enough to need all way stop signs.

Takeaway question – should we ever use traffic cops or just go to all way stop control?


Need to calculate roundabout capacity? Check out our roundabout capacity analysis spreadsheet. An easy-to-use spreadsheet that can easily be adjusted for standard defaults, such as lane utilization, critical headway or follow-up headway, according to local factors.


  • It’s worth noting that they constructed a Concentric Single-Lane roundabout, the type that has separation between entry and exit points rather than a unified conflict point. The drivers, taking the fastest path in the absence of a truck apron, drove in a pattern that began to resemble a crossing roundabout.

    The difference in gap selection of crossing versus joining means that, had they designed a MUTCD-compliant roundabout (without separation between entry and exit conflict points), the intersection probably would have performed even better. Under the MUTCD design, drivers also wouldn’t be decelerating in order to exit the intersection, which causes greater “friction” within the intersection due to braking, which also would have increased the throughput.

  • We have a roundabout being built very near our home which I will be using to get to work. WE CANNOT WAIT!!! The intersection has long been a pain, as I could use the intersection on the way TO work, but couldn’t use it on the way back, forcing us to either use business’ driveway (badly in need of repair) or go way out of our way to get home. This is the answer to many prayers!

  • Kuching has a network of two lane urban arterial roads meeting at intersections but at the intersection the road is widened as much as possible to 3 or usually 4 lanes. Traffic light phases are strictly one way goes, 3 ways stop. When the green phase starts the 2 outer lanes turn right and left and the middle two lanes proceed straight across. This means that lots of waiting traffic gets away in the first 20 seconds about 25 vehicles despite the fact they they are from a standing start. After 20 seconds the normal flow rate proceeds through.
    Not sure if Kuching’s layout is unusual.
    It does mean the experiment doesn’t work for Kuching because there are only 2 lanes at the lights.

  • In theory, roundabouts might be safer due to lower speeds, etc…However, where I live, rarely does anyone entering the roundabout ever look left!!! I will be in already and someone will come up to their entrance and literally never even turn their head, or slow down. So, theoretically they could be safer…if you don’t have inattentive, distracted, self-absorbed drivers.

  • What is the difference between a traffic circle and a roundabout? NJ has been trying to get rid of traffic circles for years (example the Ledgewood circle); this “roundabout” looks just like a traffic circle to me.

  • A modern roundabout has yield on entry control and typically has a design speed of 15 to 20 mph at all points in the roundabout. Traffic circles do not follow these rules.

  • {"email":"Email address invalid","url":"Website address invalid","required":"Required field missing"}
    Mike Spack

    My mission is to help traffic engineers, transportation planners, and other transportation professionals improve our world.

    Get these blog posts sent to your email! Sign up below.