Intellectual traffic jam

Imagine you’re on a highway with two lanes in each direction. Two cars are traveling side-by-side at exactly the speed limit. No one can pass, and so the cars immediately behind the lead pair go a little slower than the speed limit in order to maintain a safe distance. This process cascades until traffic slows down to a crawl miles behind the pair of cars responsible for the traffic jam.

Something similar can happen in organizations. Suppose the person at the top of a company is afraid to hire anyone smarter than himself. He wants to hire people who are talented, but not quite as talented as he is. If each of his subordinates follows his lead, the result can be a lack of talent at the bottom of the org chart.

13 thoughts on “Intellectual traffic jam

  1. “the cars immediately behind the lead pair go a little slower than the speed limit in order to maintain a safe distance”

    That’s a flawed assumption. After driving a little bit slower for a short period of time, the second pair of cars can speed up to drive at the same speed as the leading pair. The distance, providing everyone’s driving at a constant speed, is going to remain the same.

    Similarly, talent//smartness is not unidirectional. It’s not a vector it’s a field.

  2. I know it’s just a metaphor, and I shouldn’t pick at it too much, but:
    Why do the cars behind have to go slower? As they approach the two lead cars that cruise abreast, they will have to slow to the speed limit at a safe following distance. They might initially overcompensate. But ultimately, the system ought to stabilize to a point where everyone’s going the speed limit (given a sufficiently long highway, such that road capacity doesn’t become the limiting factor).

  3. At first it may not seem that the cars behind the lead pair should have to travel slower, but this has been empirically verified. Policemen would travel in pairs at the speed limit in hopes of making everyone behind them drive the speed limit. This resulted in disastrous traffic jams. Worse, the policemen causing the traffic jams would not immediately realize what was happening since all the trouble was downstream.

    Drivers speed up and slow down for various reasons over time. Say someone’s speed varies 10 mph. On an open highway they can average 55 by driving between 50 and 60. But if someone in front of them is driving a constant 55, they will have to slow down to 50 in order to be able to maintain their usual variability.

  4. Bill the Lizard

    This reminds me of the old rumor that went around claiming that Google’s hiring strategy was that any new hire had to be smarter than the average of all employees currently at the company. The flaw in this strategy is revealed if you follow it all the way back to the founders. Following this policy, either Larry or Sergei would, by induction, necessarily be the dumbest guy at Google. I highly doubt this.

    I suppose they could have adopted the policy after they had enough employees to make a statistically significant random sample, but I prefer to go on believing that I’ve debunked an urban myth. :)

  5. People’s speeds as a reaction to the car in front are oscillatory and damped according to the density of traffic. If the next car is a mile away, there’s no traffic jam. On a Bank Holiday weekend, when everyone and their grandmother is going for a long weekend, the density is high and cars are close together. In this circumstance people overcompensate for the actions of the car in front. You can see the ultimate effect of this in the video ScottS-M linked to.

    One way of clearing these kinds of traffic jams is to gradually slow the traffic behind the traffic jam – 70mph to 60mph to 50mph etc. This way the speeds are less reactive and a gap opens up ahead of this traffic in which the traffic density is lower and the compression wave can dissipate.

    One day, we’ll drive onto the motorway and our cars will enter a kind of following mode where they synchronise their speed with nearby cars and drive a few feet apart. This is highly fuel-efficient and negates all the above problems entirely.

    So, drivers don’t just end up driving very slowly as a stable state, as the OP suggests, but oscillate. In high density traffic, they can come to a temporary stop.

    Another explanation of mediocrity in large corporations is this: if you’re good at you’re job, you get promoted. Eventually, you’re in a job you’re only mediocre at, so you stop getting promoted. Thus everyone gets promoted to mediocrity. Combine this with the depressive effect an unhappy and bad manager has on his team members, and you find that frequently an issue arises that puts people out of their depth from the top down, and an avalanche of blame richochets down to the poor guys at the bottom. Enough of this and everyone keeps their head below the firing line, careful not to look clever.

  6. I was always amused to compare the behavior of cars on a freeway with gas molecules in a tube — the reactions to crowding are opposite.

    Long ago at an awesome conference I saw a spectacular presentation by some guys at Los Alalmos. They managed to simulate all the traffic in the Dallas / Ft. Worth metroplex, down to individual cars. Not only that, the cars belonged to specific people who lived in specific places, commuted to work at particular places, ran errands to different places, etc. all according to their own schedules and needs. They used census and survey data to get the basis for the underlying distributions. The cars all drove on the “real” roads, negotiating curves, mergers, traffic controls, intersections, etc. For modeling purposes they required the cars never to touch, and if the resulting braking acceleration was too high when cars neared each other, an accident was delacred and handled appropriately.

    They turned the virtual people loose and watched. They found very typical traffic patterns, even to the point of re-creating the backwards propogating delay that persists long after an impediment has cleared.

    If that’s not cool enough, they made a “real time” aerial view of the traffic at night, which was incredibly convincing.

    Finally they simulated a large scale disaster such as would be expected with a terrorist attack, or an impending one which is known. Traffic ground to a halt as you would expect. The take home lesson was that hardly anyone would be able to get to safety.

    On a happier note, they pointed out that the model could be used to assess proposed changes to the roadways, to see what the effect on traffic would really be. Traffic changes are often non-intuitive.

  7. A flawed metaphor perhaps, but I see what JC is getting at. If the leader(s) for some reason cannot bear to be passed by those that are more capable, the intellectual traffic jam can occur. This may also drive those smarter people to leave for other jobs, or possibly starting their own companies.
    There is a certain implied insecurity in the leaders that are deliberately “slowing down” traffic, but this doesn’t have to be the case. For instance, a leader might be so egotistical as to think they ARE the smartest person, and so allow themselves to be pushed along at faster speeds.

  8. Nice metaphor

    I real life tough, the cars behind fund a new company and pass the slow one. Well, at least if you’re in not in communist country (if such a thing still exists)


  9. I agree, this is not how traffic jams happen. Basically all cars on a road can travel at full speed (here, as fast as the two cars at the “front”) providing that cars are spaced at least the minimum following distance apart. At some point the density of cars on the road causes people to have to slow down to preserve a safe following distance. Minimum following distance then decreases with decreased speed, but at some point of increased loading the speed of traffic has to decrease again to satisfy this new safe following distance. Cars take up a finite amount of space on the roadways too, so adding more cars decreases following distances faster than just dividing inter-car distance by number of cars. Anyway at some point of loading, there is a fast phase change between full-speed and traffic jam states.

  10. “At first it may not seem that the cars behind the lead pair should have to travel slower, but this has been empirically verified.”

    This sounds like you’re suggesting that demonstrating that people do something is equivalent to demonstrating that it’s necessary. Of course, it isn’t. There are probably some really interesting counterexamples.

  11. Perhaps I should have said that it’s not obvious that the cars behind the lead pair have to travel much slower. They cannot travel faster than the lead pair. The could theoretically travel the same speed, but people don’t have the control to do this exactly, unlike, for example, railroad cars all moving at the same speed. So the only room for variation is down. And apparently the amount of decrease in speed is substantial.

  12. I would say there’s three cases to consider here:

    In the first case, a car is limited by a vehicle in front. Here, the vehicle in front provides the limiting speed for the vehicle behind. And if the average speed reduction of the car behind is epsilon, then when traffic is backed up so that n cars are following, the limiting speed is reduced by approximately n * epsilon at the end of the pack.

    In the second case, a car is so far behind that there’s no constraints imposed by the car in front. Here, we ignore the car in front.

    Finally, we have cases where both cars are involved, but lane changing or whatever is possible. Here, the following car’s speed is constrained but only temporarily.

    Clearly, in the context of a traffic jam, the first case characterizes the situation.

  13. Funny that everyone hyper-focused on the traffic hypothesis, only minimally addressing its applicability to the problem of effete leadership habits in the postmodern workplace, while almost entirely leaving out the actual problem looking for a solution (empowering real problem solving talent in an organization).

    As the workplace becomes more virtualized, and collaborative problem solving transactions become more asynchronous and indelible, it will be a lot easier for everyone, and the system itself, to notice and promote its real problem solvers.

    It will be interesting to see what happens at Google after it’s had a chance to figure out that Marissa’s gone. Larry & Sergey seem more like Mascots than anything else. What are the odds that posse’ll go down in history with The Lotus Corporation?

Comments are closed.