i'm not really sure i get the whole predictions market and wisdom of crowds thing.
Oh, don't get me wrong, i completely understand the value of these things and won't deny the fact that they do actually forecast trends and results. It's just that i'm not really certain that i understand all the elements that go into making the market.
Fundementally, a prediction market works on the Wisdom of Crowds concept which says, if you ask a whole bunch of people a specific question (say, the number of jellybeans in a jar) the average answer of the group will be more accurate than any individual answer. Tonight, i heard a bunch of folks talking about creating effective stock futures markets that try to forecast trends and development, often with highly correct results.
The problem i have can best be described as "The Limbaugh Effect". See, my problem with these sorts of things is that sure, ask a bunch of individuals their ideas and you'll get an accurate statement. However groups of individuals have a very different dynamic.
Anyone who's been in a meeting has probably seen this in play. You get one or two individuals who are, shall we say, more vocal than the rest of the crowd. They argue and debate, quite often in the open about the question at hand. They influence not only themselves, but others. Keep them in position long enough and they stay influencers.
i asked the board "How do you counter the loud voices?" Their answer: "You don't". Apparently, their belief is that you have enough loud voices that balance out and you effectively have a smaller market group you're sampling against. i'm not so sure. In effect, what you've got is a popularity contest between the various loud voices.
Then there's the concept of incentives. These are efforts that prediction markets do in order to make crappy questions interesting. Kind of like gambling on a sporting event where you stand the chance to win a buck if you're right. Again, that strikes me as being bad. It's not asking folks what they think, it's asking them which decision will win. (Are you going to decide against what you believe to be the popular decision if it will cost you money, even if you feel that your answer is correct?)
The funniest part is that one of the opening presenters asked "So, what would happen if you ran the country via a prediction market?" You know, have a system put in place where people had a real incentive in order to properly guess the outcome of a national poll in order to determine the fate of the country. They were quite convinced that such a system could be put in place and be very successful.
You know, rather than the whole "election" thing we do every couple of years.
Yeah, that system just sucks compared to what they're proposing.
Don't get me wrong. Like i said, if you ask a group of anonymous individuals their best educated estimate about an outcome, i'm quite certain that it will be correct. The trick is, getting the data for that anonymous group right.
Prediction markets are only as effective as the people listening to them.
That means that posing the proper question and answers or "futures", and allowing the artificial market to flux and be influenced is what makes these things interesting. Yes, you get core influencers, yes, the most effective thing you can do is to isolate the individuals. Maybe providing a reward or punishment may provide additional interest in the markets remaining active, but ultimately it boils down to one, key thing.
"Ask People What They Think And They'll Tell You."
The rest is all window dressing.
