4 hours ago
1
Examine, a free weekly newsletter covering science with a sceptical, evidence-based eye, is sent every Tuesday. You’re reading an excerpt – sign up to get the whole newsletter in your inbox.
I am a terrible transit planner. I slavishly follow Google Maps’ directions. And I always assume my journey will take no longer than it indicates – forgetting traffic, parking or just sheer bad luck.
So I often find myself running a bit late and am taken by the urge to speed to try to make the time back. Not speed speed, mind you, but just to nudge the speedometer needle maybe a kilometre or two past its legal apex.
It feels like driving faster will get me to my destination significantly earlier. Same distance, increased speed – it’s just simple physics.
“It makes sense – I go faster, I’m going to get there quicker,” says Stephen Greaves, a professor in transport management at the University of Sydney.
But, like many things involving the brain, that feeling is wrong. “Perception,” says Greaves, “tends to be different to reality.”
Speeding against the curve
Why do trips always seem to take longer than we predict? For the same reason rail and road projects tend to be much more expensive and take a lot longer to deliver: the human mind has a systematic bias to underestimate how long it will take us to do things.
Known as the “planning fallacy”, this seems to be driven by an unrealistic optimism we have about ourselves. When we’re asked to estimate how long others will take to do something, we tend to be much more pessimistic.
If we’re chronically underestimating how long it will take us to drive somewhere, it stands to reason many of us will find ourselves running late mid-trip. And then the temptation to speed kicks in – because surely travelling faster will get us there sooner.
We experience our brains as rational calculating machines, but this is not really how they work. In reality, they are always hunting for the shortest, easiest way to make a calculation – known as a heuristic. They are meant to be “good-enough” solutions, but sometimes they are seriously flawed.
The problem of calculating speed over distance is a classic case of using a good-enough method to get the answer – and getting it wrong.
To understand the problem, imagine you’re a local citizen being asked to evaluate two road infrastructure projects. Both projects cover 100-kilometre stretches of road.
The first improvement, widening a highway, will increase the average speed for your trip from 70km/h to 110km/h.
The second improvement is smaller in scope – fixing potholes in a local road. It will increase average speed from 30km/h to 40km/h.
Which improvement saves more time? Most people will opt for improving the highway because it increases average speed by 40km/h.
Yet, it is the second improvement that brings the greatest reduction in travel times by 50 minutes over the 100 kilometres. Option A reduces travel times by only 30 minutes.
Our brains tend to think there is a linear relationship between speed and time taken to get somewhere. But the actual relationship is curved.
The Y axis of this graphic is time taken to go 10 kilometres, in minutes. The X axis is speed in km/h. You can see the dramatic benefits from going from 10km/h to 20, versus the tiny benefit from accelerating from 110km/h to 120km/h.
At 20km/h, it takes one hour to travel 20 kilometres. At 30km/h, it takes 30 minutes.
At 60km/h, it takes 20 minutes. At 70km/h it takes 17 minutes. The speed increase is the same – 10 km/h – but you save only three minutes versus 30.
Professor Eyal Peer, who has studied the phenomenon, has proposed a “paceometer” to replace the speedometer, showing both how fast your car is travelling and how long it will take you to travel 10 kilometres. It would dramatically change the way you think about different speeds.
Don’t believe the maths? Greaves has tested this empirically.
In 2009, he put GPS trackers on 106 cars and tracked their speed and distance travelled across Sydney for five weeks.
His drivers sped quite a lot – 9 per cent of the time driving was spent above the speed limit. Despite that, the average driver saved just 26 seconds a day from speeding – equivalent to three minutes a week.
Even the most aggressive driver in Greaves’ study, who spent 44 per cent of the distance they travelled above the speed limit, managed to save just two minutes a day in travel time.
Two minutes is about the length of time for a traffic light to cycle. Get one bad red, and all your savings are eliminated.
“When you do the numbers, you go, ‘My God, it’s only saving 30 seconds a day on average,’” says Greaves. “Speeding is saving the average motorist 30 seconds a day. But the risk of a severe accident is going up exponentially.”
Why? Part of the reason is many of the best opportunities to speed are on highways, where you’re already driving quite fast – and as we’ve seen, the faster you are going, the less value there is in going even faster.
Another explanation: drivers in Greaves’s study spent a very large amount of their trip time stationary at traffic lights. “Over a typical 10-kilometre car trip, 80 per cent of that time you can’t speed,” he says.
Why do we struggle with this concept? Maybe because much of our mathematics education at school focuses on linear relationships and rarely on non-linear ones.
Known as the “illusion of linearity”, this bias shows up repeatedly among mathematics students of all ages – even if you try to teach them to avoid it. Linear models just seem intuitive to human brains.
Driving faster may not save you much time, but it does increase your risk of crashing (and of getting a speeding ticket).
It is difficult to perfectly quantify this risk as we don’t have enough data on low-speed crashes, says Professor Jake Olivier, deputy director of the Transport and Road Safety Research Centre. “But we do know that higher speed in a crash has more kinetic energy, and serious or fatal injury can occur when that energy is transferred to a person.”
A 2019 meta-analysis he co-authored found every 1km/h increase in speed led to an 11 per cent increase in the odds of a pedestrian fatality. At 37km/h, the real risk of pedestrian death is just 10 per cent. This relationship is also non-linear: small increases in speed above 30km/h dramatically increase the risk of pedestrian death.
The Examine newsletter explains and analyses science with a rigorous focus on the evidence. Sign up to get it each week.
