Thursday, November 18, 2010

A bad week on New Zealand’s roads.

Five cyclists were killed by cars this week on New Zealand roads. That’s a massive number for such a small country and as you can imagine it’s created a bit of a storm in the media. At the eye of this storm has been the question: how do you stop the collisions?

Between the cyclists’ lobby groups, the politicians, the police and readers’ online comments there’s been a swathe of suggestions. Some like increasing driver and cyclist training make a bit of sense, others like licensing cyclists and forcing them to pay registration fees (with a view to reducing the number of cyclists on the roads) seem somewhat less sensible.

The one idea that’s been conspicuously absent from the debate is that perhaps these cyclist fatalities are simply ‘system accidents’; unusual and undesirable, but statistically inevitable behaviour in a complicated tightly coupled system. The term, originally coined by sociologist Charles Perrow, is most frequently used in the context of aviation – more specifically aviation crashes.

Think about it like this; your average 747 has somewhere in the order of six million parts. If for the sake of argument we average it out at ten possible states per-part; working, worn, bent, broken, intermittently failing etc. Then you end up with ten to the power of six million as the number of possible states for your 747 a few of which will cause a crash. It’s a really mind bogglingly complicated system - and that’s just the mechanical side of it!

Likewise commercial aircraft have time-dependent processes (you’ve got to land before you run out of fuel) rigidly ordered processes (you really should put the undercarriage down before you land) only one path to a successful outcome (the runway is your only realistic option as a place to land) and very little slack (the ground is really, really unforgiving when you’re travelling at 300km/h). It’s what’s called a tightly coupled system.

When you consider it, it’s not really surprising that despite the best efforts of some of the most intelligent minds on the planet, these planes do occasionally fall out of the sky. For a great many years crash investigators would look for a single cause that caused the crash (usually blaming the pilot) but thanks in part to Perrow’s theories these days investigators tend to focus on the system as a whole.

Now, cast your mind back to cycling. Again it’s a mind boggling array of different interacting parts; every car that passes a cyclist is its own complex system; the road itself is a fantastically complicated mix of cars, trucks, busses, pedestrians, signs, potholes and puddles; and the drivers and cyclists themselves are subject to an almost infinite array of mental and physical states... about the only simple thing in the equation are the bikes.

And don’t take this the wrong way, but the average driver on New Zealand’s roads possibly doesn’t take their responsibilities as seriously as the first officer on the flight deck – I doubt you’d see them texting or doing their hair while taking off.

As far as room for error goes, it doesn’t get much closer than cycling on the road. Each car that passes a cyclist is a ton or more of hardened steel passing at high speed within a metre of an unprotected body that’s looking the opposite direction. If it were a workplace the Department of Labour would shut the place down pronto! At that speed and that proximity, it only takes a small failure in any one of those systems – a sneeze, a blowout, an unexpected pothole – to trigger a series of events that ends in a fatality. Sadly in the last week we’ve seen it happen five times.

Now don’t get the wrong idea, I’m not saying that cycling fatalities are simply a fact of life and we should get used to them. The point is that just like in aviation, in looking for solutions to the problem we probably shouldn’t worry too much about looking for a single cause or adding more checks and balances – stiffer penalties, better signage and more legislation.

The logical answer seems to be to focus on making the whole system simpler by reducing the number of interactions between sub systems, by separating bikes and cars into separate lanes for instance or de-coupling the system by reducing speed limits (at least in built up areas) to give people a bit more time to react.

The question of how to make cycling safer does sadly bring up another aviation concept – the catch 22. In countries like the Netherlands where the average person cycles 2.5 kilometres per day, the death rate is about 1.4 people per hundred-million kilometres cycled. In the USA where the average person only cycles 0.1km per day the death rate is almost 27 times higher at 37.5!

It seems that more cyclists makes for more political clout, which produces more bike paths, more driver training, lower speed limits – and thus makes cycling safer. But when getting into the saddle means taking your life in your hands, who’d be mad enough to ride a bike.