One of the times that this technology is most crucial is when racing at a brand new circuit, or one that has been changed since the previous year. For the first race at Fuji in 2007, engineers used computer representations of the circuit to figure out the best racing line before the cars had even been shipped to Japan. The figures from these simulations are so accurate that Willy Rampf, then BMW Sauber’s Technical Director, said at the time, “our aim is to ensure that the lap times and top speed calculated using the simulation software do not vary by more than one per cent from the actual values.”
Historically, Formula 1 has always been a breeding ground for technologies that eventually inform the design of cars that the rest of us will drive. Now that the speed kings have moved into advanced computer simulations, other industries might well start listening to them too.
Companies like Predicted Future claim they can predict events purely by analysing freely available data from the Internet. Another one, called AmalgaMood, even offers a ‘mood index’ that they claim can predict stock market downturns. Since the wide adoption of services like Twitter, people have been broadcasting their opinions and activities to the world. The potential for harnessing this sort of massive data is something that has already interested academics. A 2010 study from Carnegie Mellon University compared billions of tweets with traditional polls on consumer confidence and political opinion. They discovered that the two correlated 80% of the time.
This year, American TV network NBC’s poll suggested Rick Santorum would take fifth place in the Iowa caucus. He did substantially better than that, winning by a small margin. An analysis of tweets carried out after the fact showed that the substantial upsurge in his popularity was actually predictable.
Businesses often spend huge amounts of money on strategy and prediction. Supermarkets already use historical sales data to manage demand forecasting, but they can only look at the data their customers leave behind. If they could analyse social data to predict live demand across all consumers in a local area, they could run tightly focussed campaigns and offers in order to win business.
The retail sector has also used focus groups for decades, and for at least the past six years has trialled new products by having test-shoppers walk round virtual stores. A company in Portsmouth, called Path Intelligence, has for a while offered a product that can track shoppers around a venue based on their unique mobile phone transmissions. Leaving aside the obvious privacy concerns—which have recently seen the system hounded out of malls in the USA—such data can provide an insight into how real people use a shopping space, allowing retailers to lay out shops to the maximum advantage.
While business has long gathered information to help it plan, the advent of freely available social data allows it to map existing knowledge onto sentiment. Similarly, based on existing data about how crowds of people act in any given situation, infinite computing power could calculate endless extrapolations of how behaviour would be altered if one or more variables in an environment changed. This combination of historical intelligence with live social media could reach far beyond commercial concerns to predict, for example, the prevailing mood of a group of protestors and how different policing actions would be received. If the predictions could be computed in real-time, as in Formula 1, then future crowd control could look very different.
With endless streams of data being generated by increasing numbers of people, and limits on cloud computing power becoming largely theoretical, a perfect storm of infinite data and infinite computations is brewing. The magic will be made in understanding how to apply those calculations to our world, while not forgetting that, just as Lewis Hamilton can fly off into the gravel, humans can sometimes depart from the predictable in spectacular ways.