It’s one of the little quirks of our postmodern wonderland that slight imperfections in a quasi-utopian service are complained about far more than the absence of the same thing: you’d rather no WiFi than slow WiFi, that kind of thing.
In a similar vein, until quite recently all the humble bus user had to go on when waiting on their chariot were timetables and some educated guesswork. How quaint it must have felt. In the last few years, though, the development of various breeds of real-time software means you can see your bus’s ETA down to the minute, via apps and on-street displays.
That’s the idea, anyway.
Dublin’s version of this, the Real Time Passenger Information system (RTPI), was launched in 2011, supplying travel information to a network responsible for 128m passenger journeys in 2016. And it has a bit of a reputation.
“Why is my sister called Chaos?”
“To remind us of all the unpredictability in life.”
“What abou-“
“That’s enough, Dublin Bus Real Time App.”— Carl Kinsella (@TVsCarlKinsella) November 6, 2017
Every commuter has a story about buses supposedly coming closer before receding into the distance, or showing up on the displays, only to vanish entirely. “That thing’s been saying ‘6 minutes’ for a quarter hour now” sits alongside “shite weather, isn’t it?” in the annals of Irish small-talk. Different countries have their own version of the same popular discontent – some a little better, some a little worse.
the workers at Dublin Bus HQ deciding what false time estimate to put on their app next pic.twitter.com/n2LQUADShQ
— Jack (@jackdoylex) January 21, 2018
Basically, RTPI works by inserting a GPS signal into a bus and working out how far it is from each subsequent stop along the route. Jeremy Ryan, head of public transport contracts with Ireland’s National Transport Authority, says that the system is fed with regularly-updated ‘profiles’ of how long a bus should take to get from stop to stop, with different profiles for different days and times. These profiles are aware of normal traffic and pre-planned roadworks, but can’t know about ad-hoc things like heavy rain or emergencies, which leaves the predictions a little sticky.
The NTA carries out quarterly surveys to assess the system’s accuracy. According to this research, about 97 per cent of Dublin buses arrive within three minutes of the scheduled time, which the NTA and Dublin Bus say is “well above industry norms”. Reassuringly for people convinced that their own area is cursed with malevolent software, there apparently aren’t significant variations between routes, directions, or times of day.
Approaching these assertions with the swivelled eye of healthy scepticism, I decided to engage in a bit of citizen’s quality assessment. I walked to some strategically-placed bus stops around central Dublin, waited for a bus’s expected arrival time to tick over to “5 minutes”, and measured how long it took to roll up beside me.
Not having all day, I limited myself to clusters in Phibsboro and Rathmines, 2-3km north and south of the city centre respectively, and to the city centre itself from O’Connell Street to South Great George’s Street.
And what do we find? Well, probably unsurprisingly, on the whole the RTPI system does indeed appear to be remarkably accurate. Averaged across 62 buses, the average time it took for one to arrive after the ticker moved to “5 minutes” was a lean 5:58. Bearing in mind that nowhere is it claimed that “5 minutes” should be taken to mean 5:00 exactly, keeping the average error down to less than 20 per cent is not too shabby a performance.
This makes all kinds of sense. A system with much more of an inbuilt error would be nigh-pointless to continue using from a passenger’s perspective. And it’s easy to see how a system with even a 3 per cent error would develop a negative reputation, considering how many people it stands to annoy every time there’s a slip-up.
That said, at different points of my experiment, “5 minutes” could have meant either 2:32 or 11:21. If you arrived at a stop proclaiming “5 minutes” on a given day last week, it seems there was about a 1 in 4 chance of the bus being more than 2 minutes early or late. Insofar as there was any variation between areas, buses heading into the city centre from the south side averaged 6:41, though there were hardly enough trials here to be statistically significant.
In sum however, we’d probably be advised to give the software some credit. For all I know, it may well have a conception of time divorced entirely from our idea of reality – but for the most part, when RTPI tells you your bus is 5 minutes away, it’s not exactly lying to you.
