How do routefinders find their routes?

  • 14 April 2011
  • From the section Magazine
  • comments
Image caption Computers now often decide the way

Many now rely on satnav or online routefinders for directions - instead of paper maps - but how do these gizmos find their way?

Once, you gazed at the map, remembered that cursed bridge on the old A30 near Bodmin where lorries got wedged, thought of a number, doubled it, gave up, guessed, set off at 4am, whatever.

Now, you consult an online routefinder.

Or maybe eight, as I did. I wouldn't usually, honest. It's like having eight argumentative backseat drivers, as you'll see.

So how far is it, anyway? And how long does it take? The occasion was another bank holiday, the journey was Hertfordshire to the South Devon coast near Kingsbridge, and here are the results:

Is that a bigger spread than you'd expect? The distances are 246, 244, 243, 242, 221, 221, 220, and 220 miles. The durations vary by more than an hour. But the explanation is not that some are adjusted for traffic conditions.

Four routefinders recommend the M4 down to Devon, the other four the A303. When trying a few different times of travel on the TomTom - which is one of those that says it updates according to traffic conditions during the day - it varied by a maximum of 20 minutes from quickest to slowest in either direction.

I had expected more. In the event, we passed a life-sapping tailback on the M5 in the other direction that would add who-knows-how-much suppressed road rage.

So how do they do it? The one I spoke to was a bit cagey about this in case the competition would be listening. But the basics are that they first assume you drive legally, count the distance, make standard calculations for how much time a roundabout or junction adds, then check samples of these with real drives.

There's apparently no encyclopaedia of routes from every destination to every other. Rather, they look for the shortest way to the trunk road network and then refer to a database of main routes, before finding the shortest reasonable - ie not down a farm track - connection at the other end.

So we might expect variations to tend to be proportionately bigger where less of the route is on a trunk road.

But here's a question that moves the problem of measurement from one about miles to one about you. Do you choose the route finder that offers the quickest journey, the slowest, or one in the middle? Are you a travel optimist or pessimist?

If you think it possible that one of them knows a better route, it might make sense to choose the fastest. But how likely is that? Makes you wonder why they don't put a range of uncertainty around the journey time.

The distance is what it is, the journey will take what it takes. As they used to say about the timetable for certain continental rail networks - you leave when you leave and arrive when you arrive. This doesn't change just because someone thinks otherwise.

But I wonder if the hope that it might is common in many areas of life. For example, do you choose who values your house, audits your company or sets your exams because you think they will give you the most favourable verdict? Or the most accurate?

In other words, do we go for the one that we think will tell us what we want to hear - and stuff the facts?

Come to think of it, you do meet people who like to say they did the journey faster than they really did - yeah, my driving's that good. Or slower - see how I struggled to get here. See what you put me through?

It is what it is, but measurement is something else. Measurement, as we all know, is also about the measurer.