Traffic Sense

I’ve discovered most city planners do a good job, but those who control the city planners i.e. mayors and councils, well, they don’t. For that matter, a lot of city planners are “old school” and are dumbfounded at the remarkable safety and efficiency of multi-lane traffic circles.

As you watch the following video, I’d like you to know that excellent traffic managers not only study and understand interchanges, but they also understand how traffic flows between various traffic junctions. They approach traffic not as a collection of nodes, but as a system, where decisions made about how one node (intersection) should operate affects other nodes.

I’ve been a systems analyst since that fateful date during the winter of 1989, when I worked hand-in-hand with a team from Booz-Allen at Louisiana Power & Light helping the utility to get a better handle on its inventory management.

Back to traffic: How does your city fare? Let’s find out:

Quick Grade Sheet:

1. Is your city relatively free of stoplights? Or does the city slap a stoplight at every dangerous intersection created by their own poor planning?

2. Does your city make regular use of traffic circles aka roundabouts to facilitate the unimpeded flow of traffic? Or have they resigned themselves to the very wrong idea that if half the traffic is moving, the other half has to be stopped?

3. In those areas where stop lights are an absolute must, do they employ reactive sensors requiring vehicles to stop before the light turns green, or do they provide proactive timing by networking with other traffic lights to facilitate long flows of uninterrupted traffic by creating virtual thoroughfares (green light before you and others in your “packet” arrive)?

4. Are your city traffic managers of the mindset that traffic is something that must be controlled in order to minimize problems or do they see traffic as a customer whose needs should be facilitated?

Traffic problems begin with incorrect, if not flat out wrong mindsets.

Traffic solutions begin with a keen desire to learn the myriads of ways cities can ditch traffic lights and in favor of Smooth Flow Solutions.

One city laying in a large valley, for example, realized more than 80% of its traffic went north and south. They designated even numbered streets using the “Even Up” approach as north-only lanes, while designating odd streets as south-only lanes.

Both stop signs and stop lights on these streets were removed.

The city was about five miles long and a mile wide, and designated seven of the highest-capacity east-west streets as thoroughfares. The intersections of the east-west thoroughfares and and the north-south streets were given stoplights.

But they didn’t stop there.

They designed the stoplight timing to allow for north-south packets of traffic to move continuously. The north-south lights would only change red after a large packet had transited, clearing the way for east-west traffic along the thoroughfares. Timing allowed east-west flow until the next north-south packet was scheduled to arrive.

But they still weren’t done…

They built roundabouts (traffic circles) to handle multi-directional edge traffic.

Instead of trying to build a single freeway intersection to handle two sets of traffic flowing both ways, they broke it up, interacting with the east-west freeway at the north end of the city in two places, first with one of the north-flow thoroughfares, and again with one of the south-flow thoroughfares not one but three blocks away. As a result, they eliminated having to deal with crossing traffic altogether, except for the two trunk lines, which required bridges, anyway. No stoplights at either of the interchanges, but for a total cost that was half that of a single cloverleaf design.

Result: Average commutes were cut in half, even after traffic density doubled.


Four-way stop signs are just about the absolute WORST way of managing traffic. They are HORRENDOUSLY STUPID in ALL situations.

Two-way stop signs can be used in low-density traffic grids to provide some relief. Stop signs can alternate every other street for nearly equal flows in all cardinal directions, or you can use two 2:1, 3:1, 4:1, etc. in one axis, at the expense of the other, to give significant preferential treatment. You can also checkerboard supersets of intersections in one direction while…

Having read and thoughtfully considered all of the above, let’s watch that video, again. After all, it’s only six minutes.

Do you notice anything different?

Here’s what I noticed:

The Stack Interchange, where traffic is never interrupted or looped, provides for the greatest flow of traffic at 1,099 vehicles per minute.

A 4-lane road without any traffic lights will achieve nearly 20% of that throughout, at 191 vehicles per minute, but it’s very dangerous. Adding traffic lights actually increases the flow to 235 or 303 vpm while greatly increasing safety.

When you introduce slip lanes, which are removed from the central part of the intersection, the throughput rises to 465 vpm.

Eons ago, roundabouts were designed such that entering traffic had the right of way. This required stops and resulted in instantly clogged circles. The traffic flow rate for such a “standard roundabout” is an abysmal 8 vpm.

Modern, multilane roundabouts, however, achieve 360 vpm — 50% greater throughput than traffic lights — by giving the circular traffic the right of way, thereby facilitating the flow of traffic through the roundabout while allowing incoming traffic to meter their own influx to maximize the traffic throughput. Using slip lanes to offload right-turn flows bumps the throughput up to 532 vpm. In fact, the Traffic Circle located in Pinehurst, NC uses slip lanes. Even though they’re adjacent to the traffic circle lane, they’re separated by white lines.

In fact, I helped design that traffic circle. 🙂 I lived there at a time when they were doing some serious reconstruction, and the newspaper contained a diagram of the local DOT’s proposed changes. Having driven all over Europe, where traffic circles abound, I instantly spotted the mistake, hopped in my car, visited with the engineer, and sketched out my proposed changes, which remain in effect to this very day:

Similarly, the U.S. Air Force Academy’s North Gate used to have a cloverleaf design, but they got rid of the eastern 270 degree loops (partial cloverleaf) and stoplights in favor of a double traffic circle (Dumbell) approach which gave them a 443 vpm flow rate, but saved a lot of concrete while simultaneously increasing local (non-Interstate) flow rates during football games, graduation, and other high-population events.

But the king of intersections remains the Stack Interchange, wit a traffic flow rate of 1,099 vpm:

The Stack Interchange allows vehicles turning right to turn right, a perfect 90 deg turn with minimal slowdown. Those going straight don’t slow down at all. They just keep on going. Vehicles turning left go through slightly more than a 90 degree turn, but not much, and again, they do so with minimal slowdown.

The greatest benefit of a stack exchange is that you can have two, three, four, eight, whatever lanes of straight traffic. You can even have two or more lanes turning right or left, as well. The point is you build it with what you need while leaving room for more lanes to be added in all directions, whatever’s needed in the future, without needing to interrupt any existing traffic. Again, done right, existing traffic should experience no slowdown as additional capacity is added in the form of more lanes.

Now, “Diverging Diamond Interchanges” might “keep you from dying,” but do they really? There are EIGHT merge points and EIGHT stoplights required for this device. Merge points and stoplights are where many accidents happen, and crossing to the left of oncoming traffic can be confusing if you’re not used to it, so I really do NOT see any safety advantage as claimed in the video below:

Despite a claimed throughput of 644 vpm, I’m just not seeing it. The greatest problem I see with the DD Interchange here in Colorado Springs, located at the intersection of I-25 and Fillmore Rd., is that the crossing stoplights aren’t on long enough to clear the center structure. It’s the same old, “We have to be fair!” so they cycle the stoplight too frequently to obtain adequate rates of flow across the Interstate. As a result, the Interstate off-ramps are often backed up as much as they were before. To make matters even worse, city managers and/or planners put full, four-way stoplights with NO slip lanes (flow rate of 235 vpm) on either side of the interchange, thereby mucking up the works even further! See it for yourself:

Remember, a well-designed two-lane traffic circle with guarded (white line) right turn lanes (three lanes total) achieves the same throughput as a a custom roundabout with slip lanes: 532 vpm. Now, you’d have to remove both the IHOP and the Waffle House to fit it, but hey, at least the IHOP was built after the DD, so that wouldn’t have been a problem if they’d bothered to conduct proper planning.

So, don’t believe the hype! Double Diamond interchanges are not a panacea.

So what’s my bright idea for how they should have done Fillmore and I-25? A Split Dumbell with adjacent slip lanes that would have achieved in excess of 600 vpm with NO stoplights and only FOUR merge points. It also would have eliminated the stoplights on the west side of I-25!

But, hey, what do I know? I’m just a polymath who’s studied engineering, business, IT, aviation, mastering the latter three, with three degrees overall, two science, and two masters (with two concentrations), both summa cum laude, so whatever you do, don’t listen to me, as I can’t save you millions while reducing accident rates by a large factor…

But then there’s the cost factor. Obviously, a Stack Interchange is going to cost more, right? Well…

Then there’s the throughput factor. You need to achieve the minimum required throughput, right? Well, yes. But…

The appropriate way of approaching this is to find all solutions that allow you to achieve a minimum required throughput, then eliminate those that won’t physically fit, then sort the remainder by total long-term cost of ownership in an NPV (net present value) approach.

Here’s yet another video detailing some of the costs and throughputs from the previous list:

And now, for my next trick, I’ll compare these designs on a cost per vpm basis. But for that one, particularly adjusting for the time flow of money, I’ll need to be paid. 🙂

However, if you’re a city planner and want some modeling software, it’s as cheap as $30 for the basic package and $40 for the package with all the nice stuff.

Just one of many pages of very detailed options!

In closing, I’d like to say, Good Luck, City Planners and City Managers! There ARE FAR better ways of doing things!

Final one, showing many of the extremely detailed options:

Leave a Reply