Posted by Emile Servan-Schreiber on Wed, Jan 04, 2012
Prediction markets are competitive betting games designed to tap collective intelligence to predict the future.
Traditional prediction markets resemble stock markets where people can buy and sell predictions to and from each other until some real outcomes are observed and settled. However, the term is also often used to describe any betting venue that rewards participants in proportion to accuracy, precision and timeliness. Prediction markets thus provide incentives for both timely and truthful revelation of opinions.
In the two decades since the first modern prediction market began operating at the University of Iowa's Tippie College of Business, this form of collective intelligence has produced an impressive track record of accurate forecasting in fields as diverse as business, sports, movies, politics and medicine.
How accurate?
It is impossible to quantify the accuracy rate of prediction markets in general, for it of course depends on the intrinsic difficulty of each prediction. However, as the magazine MIT Technology Review once declared, "in many cases, prediction markets are astonishingly more accurate than other common forms of predictions such as polling, surveys and expert opinions." The Economist later surmized that "the most heeded futurists these days are not individuals, but prediction markets, where the informed guesswork of many is consolidated into hard probability."
Specifically, the data show that:
It is often said that prediction markets are accurate because they require that people "put their money where their mouth is". However, it has been shown that the predictions of play-money markets can be just as accurate as those of real-money markets... So the real story is obviously more complex and interesting. In fact, prediction markets derive their predictive powers from three sources:
1) The first source is cognitive diversity. In a nutshell, when dealing with complex issues involving many variables or moving parts, no one can claim to have a complete model or theory from which to make fail-safe predictions. More likely everyone has a partial understanding of the situation, further clouded by his own biases. But when all these partial, biased models are put together, a wonderful thing happens: knowledge accumulates, gaps get filled, while the various biases cancel each other. The group's collective model is better and more complete than any individual model. Prediction markets attract diversity like powerful magnets, because anyone with a model that disagrees with the current consensus has a profit motive to participate in the market.
2) The second source of market accuracy is that the process encourages the voicing of informed contrarian opinions rather than calculated conformity or respectful consensus. Indeed, the only possibility for profit lies in disagreeing with the consensus, publicly. This makes sure that all informed points of views are included and aggregated. Importantly, the possibility of financial loss - whether the currency is real or virtual - also discourages and penalizes the voicing of non-informed opinions.
3) The third source of market power is that framing predictions as wagers makes people think different, literally: Brain imaging studies show that when it contemplates a gamble, the brain becomes more risk averse and tunes out the emotional signals that might interfere with cognitive performance. In short, our thinking becomes more objective and our judgements less clouded by our passions and preferences.
Lumenogic has more than a decade of experience designing prediction markets to help companies improve forecasting, gain customer insights, create idea markets to power social innovation, and enhance portfolio management. Contact us to explore the possibilities!
Posted by Emile Servan-Schreiber on Wed, Jan 04, 2012
When John Donovan became the new CTO of AT&T in 2008, he was "led to believe we would have an innovation problem," as he recalled recently in a fascinating New York Times interview. He proceeded to gather the 16 smartest people in his department in a conference room and they bombarded him with "a million reasons" why the company could not innovate.
In response, Donovan challenged them to start innovating right then and there. Each person in the room would present 6 new ideas and everyone would vote them up or down. The top 2 would then get nailed down and funded immediately.
Today, Donovan's get-it-done approach has expanded into AT&T's Innovation Pipeline, an innovation program that, so far, has grown to involve more than 100,000 employees and to generate more than 15,000 ideas, 45 of which have been funded with "tens of millions of dollars".
The front end of The Innovation Pipeline ("TIP") is an online collective intelligence platform with a prediction market at its core. When a new idea is suggested by a participants, others can comment on it and use virtual currency to bet for or against it. Then, at the end of every quarter, the top-rated ideas are pitched to a wise council who decide on the spot which will be funded. In the end, two types of participants are rewarded with prizes, trips and other goodies: those that generated the winning ideas, and those who correctly bet on the winning ideas.
According to Donovan, TIP has become very popular with AT&T executives. One reason is that it is much cheaper to submit a problem to the eager crowd of creative employees and reward the best ideas with a few prizes or a trip to Hawaii than to spend $1 million on external consulting.
Perhaps even more importantly, TIP offers an environment where ideas are transparently evaluated by peers rather than getting killed by "some sort of corporate machinery". This, plus the fact that the company is committed to funding the ideas that come out on top gives participants the confidence and motivation they need to contribute their best.
It is a great success story for collective intelligence and prediction markets, and it shows that a CTO's role can be crucial in creating a new culture of innovation. As donovan concludes: "all we did is we digitally removed real cultural barriers by thinking through how we can use technology to solve that problem."
If you think this kind of approach might be beneficial to your own organization, let us show you how.
