Near the start of the year, I ran into an article claiming that, by slowing down in advance of traffic jams, a single driver could break up the jams, and that in particular you could turn merge congestion into smoothly flowing traffic.

Clearly, this demanded further experimentation. So I gave it a try: when I saw slow traffic ahead of me, I just took my foot off the gas, gradually slowing down until I got to where traffic was slowest. The rules of the game were that I lost if I actually came to a complete stop, and I took off points if I had to use the brakes; the goal was to change my speed as gradually as possible, in order to mute changes in traffic speed.

And it was great! My commute was much less frustrating; I was spending almost no time feeling mad at the fact that I was stuck in traffic barely moving (there’s a big difference between going 10-20 miles per hour and going 0-2 miles per hour), and I had something new to think about and keep my mind busy while driving. And it had a big effect on my gas usage: I went from having to fill up every weekend to every other weekend. (I don’t think I actually cut my gas consumption in half, but it was a noticeable difference.)

I went along like that for a couple of months, enjoying my new commuting joy. But then I remembered what the article was talking about: it didn’t just say that your commute would be more pleasant, it said that you’d be able to see miles of smoothly flowing traffic behind you. And I wasn’t seeing that.

So: what are the possible hypotheses to explain my (lack of) observation? I came up with a few:

  • The article was wrong.
  • The article was right, but I was doing the wrong thing.
  • I was having the predicted effects, I just couldn’t tell.
  • I was doing the right thing, but differences between my stretch of the highway and the author’s stretch of highway meant that the techniques wouldn’t work for me.

All of which, actually, were pretty plausible. The first two are obviously possible; as for the third, I’m driving a compact on a flat highway, which means that, on average, I don’t have to go more than about two cars away until my gaze encounters a car it can’t see over/past/through. And, as for the fourth, I only pass three exits on the way home; the first has metering lights, the third one has an extremely short merge that probably causes its own extra problems, and maybe there’s something weird about the second one that I’m not seeing. (Enough cars wanting to enter to overload any mitigation efforts?)

But the first hypothesis to investigate is that I was screwing something up. Which I was, in one obvious way: I was staying in the second lane while, if I want to have an effect on merges, I need to be in the right lane. So I did that for a while, but it wasn’t helping, at least most of the time. (It probably did help when I left enough room to let every single waiting car enter the highway, but that usually wasn’t possible.)

And then I figured out something else that I was doing wrong. What I was doing was trying to drive in such a way that I would slow down as little as possible; my algorithm for doing this was to shrink the space between myself and the car in front of me as gradually as possible, finally using it all up (except for necessary safety room) when that car was able to speed up again. In merge situations, that happens right at the merge point, so basically I was shrinking the following space down to zero right at the merge point.

The goal, however, is to set up a situation where both merging lanes are moving at a good speed, so they can merge in a high-speed flow instead of in a stop-and-go fashion. And, by minimizing both my speed and my following space at the merge point, I was working against that.

So I tried to slow down a bit earlier, to be able to enter the merge area at a higher speed and with more room in front of me. Which turns out to be really hard! My brain and right foot had been happy to start coasting earlier than they were before, but I ran into a real psychological barrier when I tried actually going so slow that I was actively holding up traffic behind me more than necessary.

I wish I had a clearer ending to this story than I do, but here’s the current state of affairs. I try to leave a little room ahead of me at the merge point, and have my speed not drop to almost zero. Sometimes I manage that; sometimes I don’t. And, sometimes when I do leave room ahead of me, the next five cars entering the highway all take this as an invitation to enter the highway, cut in front of me, and then slam on their brakes in order to avoid crashing into each other.

But sometimes I manage to sail through what had been a stalled merge point at 15 miles an hour or so, while letting a couple other cars enter the road at the same speed; it feels great when that happens. I still don’t know if it’s having any long-lasting effects, though: I can’t see far enough behind me to tell what happens to that merge point after I’ve gone by. And I’m pretty sure that the one location with a tiny merge area is incapable of staying congestion-free on its own: as far as I can tell, few enough cars use it that it empties out completely on its own every couple of minutes, which means that the backups that form there are impossible to cure without redesigning the exit. (Which sounds like a great idea to me.)

Having said that, I am convinced that there are benefits to easing off the gas when approaching a slowdown. It doesn’t slow down your commute, it makes you (or at least me) feel better, and it saves gas. So it’s a clear win, even if it doesn’t help the traffic behind me; I’m pretty sure it doesn’t hurt them, and it helps me, which is good enough.

Post Revisions:

There are no revisions for this post.