Taxis are now being guided by the same tech used in city-building games

A city-building game, no matter how wild it looks, is an economic model. Everything can be mapped. Everything is built on a substrate designed to facilitate bean counting. Everything is fundamentally knowable. There are surprises for the eye, but nothing is truly surprising.

The city-building game, in other words, is every business’s dream. A centrally controlled, fully quantifiable universe would grant considerable powers to a large corporation. That is why it’s no big surprise that the British startup Immense Simulations is using a simulation technology called Improbable to solve taxi dispatch problems. These are, in effect, two sides of the same coin.

Using Improbable, the company created a complex recreation of Manchester, England, with its tangle of roads. “Improbable’s technology allows us to run city-scale scenario simulations with a level of detail not previously possible,” says Immense Simulations CEO Robin North told MIT Technology Review. “We can represent individual travelers and vehicle operations and their interactions within a living city.”

To understand why this matters, it’s worth thinking about what traditional maps are good for and what they can and cannot do. A map—the kind you unfold on a table or over a steering wheel and then fail at folding back up—is good for big picture thinking. It can show you the way from point A to point B if you draw that line in pen. The ends of that line, mind you are somewhat rounded from being drawn in pen. They have too much ink from where your hand paused. The map will get you close enough to your real destination, but it will not solve the smaller questions about what to do in the final dozen feet of your journey—at least not unless you stick your nose right up against the map.

In most cases, that is good enough. You bumble about for the last few feet trying to figure out where your destination is, and then come up with an elegant solution. But what if you’re the head of a taxi company, say, and every second counts—for you, your drivers, and your passengers. All of a sudden, traditional approaches to mapping don’t really cut it. Small routing questions, like ensuring that a driver approaches on the correct side of the street or is quickly facing in the correct direction for their next fare, matter. In a world where everyone can draw a line in sharpie from point A to point B (or do the equivalent with cheap, dash-mounted GPS), competitive advantage can only stem from a deeper knowledge of space.

We don’t yet map space like driverless cars, but this is the transitional phase

New approaches to transit (read: Uber et al) require a different understanding of the city. They are less about the bulk of routes than systemic knowledge of urban events and microscopic details at both ends of a trip. These, incidentally, are the points where traditional forms of mapping fail and where simulations excel. As ride-share companies transform—for better and for worse—taxi dispatch into a ruthlessly efficient economic system, they need a model of the city that maps onto this outlook. Sim City is admittedly more fun than Travis Kalanick’s plans for the future, but in both cases the city is treated as a data-rich system atop which interesting infrastructure can be built.

Eventually, mapping can come to reinvent a city. Earlier this week, the GPS manufacturer Waze announced that it would cut down on left turns in LA for safety reasons. In a way, this represents a mainstream adaption of UPS dispatch’s approach of ordering three right turns instead of a left. But Waze may go further. Per CBS: “The company is now working on what could be a controversial next step, a new feature in that country to alert drivers about routes through high-crime areas.” Maps can look like destiny or they can create confirmation bias; when a map tells you to avoid a neighborhood what are the chances it will make it off the blacklist? This is all the more the case when thinking about taxi dispatching, which, at present, is really a stopgap for driverless cars. Mapping and simulating every last inch of the city turns the driver into an aspiring robot.

We don’t yet map space like driverless cars, but this is the transitional phase. It is only fitting, then, that Immense Simulations is bridging different forms of urban understanding. The city-builder isn’t quite real, but it is an urban environment. The dispatcher’s simulated city, likewise, isn’t reality as we tend to experience but it is nonetheless an urban environment. These are all realities in a sense, but it is not yet clear exactly which one we are living in.

h/t Technology Review