Wiki article
Compare versions Edit
by Hubertus Hofkirchner -- Vienna, 09 Jul 2018
The other day, at a fun event hosted by Lightspeed in Munich, I meet this fellow from an agency. He asks the obligatory small talk question: “So, what is it that you do?” Dutifully I reply: “Prediction Markets!” which results in the well-known blank stare and the little pause which always reminds me of my lamentable shortcomings as a prediction market evangelist. Ever helpful, I add: “We don't just ask people to say what they will do or buy. We ask them to bet what other people will. It makes them think much harder and gives more accurate results.”
His eyes light up. “Ah, I see how that could work very well...” I, of course, rejoice: just three sentences and my conversation partner understands our method. I give myself an imaginary pat on the shoulder for some excellent evangelising there. He ponders a little bit. “This is quite different. I assume some people are better at this than others?” Happily I accept this license to torture him with the many fantastic qualities of prediction markets: “Why yes! In fact, clients often use our method to identify the true experts in a crowd, such as knowledgeable employees or power consumers. You know, anybody can google facts but asking for what will happen in the future sifts out the true experts from the many self-proclaimed ones.”
He nods but then again starts to frown pensively. “Now that I think of it, I am not sure about that expert-finding thingy. At our agency, we are running an internal tipster tournament for the Soccer World Cup. However, the nerdy chap who’s ahead in the betting is definitely not a soccer expert, he hardly knows any of the famous players or that FIFA changed the offside rule quite a while ago.”
Now it’s my turn to stare at him blankly. After more than a decade of doing precisely such expert filtering in practice, this just does not make sense. “How can that be?” I ask, perplexed. “Well, eventually we did find him out”, he answers. “All the guy does is looking up the bookies’ odds and then he bets on the team which is ahead there.”
Ah, this solved the puzzle instantly. “Sports betting”, I explain, “is a primitive form of a prediction market. It’s really, really hard to beat a market. Your guy knows this and used the market’s collective intelligence to guide his own betting. But the experts which we want to identify are actually to be found in the sports betting market, they are those who make money there.”
My esteemed readers will not mind a bit of scientific background to this. It was the philosopher and nobel laureate Friedrich A. Hayek from Vienna, who first proposed in 1945 that market prices are but a system of communicating expert knowledge, they capture and disseminate information in the most condensed form: just one number. Hayek held that prices are all you need to know to make the best possible business decisions at any point in time.
For the next few decades mainstream economists disparaged Hayek’s theory, asserting that each single market participant would need perfect knowledge of all relevant facts if the price was to be right. Until in 1982 another nobel laureate, Vernon Smith from Witchita, Kansas, actually put the now famous “Hayek Hypothesis” to an empirical test in a laboratory. His results were consistent with the unimaginable: market prices did aggregate dispersed information of market participants, both public and private knowledge with high reliability and accuracy.
In other words, the betting market odds were all their guy needed to outperform the soccer experts in his company, even those who spent hours and hours for many years to observe and learn about the game and who may know the FIFA rulebook by heart. In fact, even a monkey can use market prices with good results but that’s a story for another day.
The good news here is that with a modern second-generation prediction market, you can avail yourself of the power of this principle for just about any business, economic or societal topic where you need the best possible information to make a real-world decision. What’s holding back the method, despite all my evangelising? I don’t know but maybe many corporate innovators are just not innovative enough to innovate how they innovate.
