I’m tardier than some in posting reports from the 2007 NY Prediction Market Conference (where I moderated a very short panel). Please forgive … and without further adieu, here’s item #1.
Content, Does it Matter to a Contract’s Price?
Robin Hanson took a question from the audience that went something like this (if I recall correctly):
Does having related content and links appear alongside a contract make any difference as to how well informed the traders are when trading that contract … and therefore, does it have any impact on the price of that contract?
Robin didn’t have much of an answer (it does seem such a specific question), but Brent Stinski, sitting near me, had some thoughts which he muttered under his breath, and at the time I didn’t quite catch all of them. I followed up with him later to get his full thinking on the issue. As the creator of MediaPredict one can expect that he does have the relevant experience to at least begin to answer this question. His thoughts were:
We feed information into all of our markets, but we haven’t had any reason to believe it skews results. On Media Predict, you can talk about a movie having great reviews or (more crucial factors) how many screens it will show on, or whether there have been any recent entries in its genre. People still predict what they want to predict.
This is “collective” intelligence, after all. In the end the editorial voice on the page is just one little voice on the internet. Traders will ignore it as quickly as they ignore other traders betting against them.
Moreover bear in mind that a trader’s only goal is to maximize his or her own portfolio, and in that situation people will be more resistant to influence and will go with their gut. (Psychologically speaking, this is probably a good thing in terms of getting better input from users, which is one reason why MP prefers markets to competitive scoring.)
A good answer. And it opens the question as to whether offering traders blogs, the ability to trade ideas, and other sorts of social networking features (as seen here) makes any difference to the accuracy of a contract’s price (although it certainly adds more content).
An area of study, perhaps …?
As always, thanks for listening.
~alex
Content schmontent …?
Loading ...
Hanson is asking the kind of question of interest to exchange operators and economists worried about market efficiency. Presumably data could be collected and results examined for bias. My intuitive sense is that Stinski is right, traders can ignore news presented by the exchange just as easily as they ignore anything else.
In fact, seeing news presented along side the trading interface may lead the trader to assume that the news is already common knowledge and therefore already embedded in the current market price. I’m more likely to trade on information found elsewhere because it is more likely to be the source of a differential advantage.
Exchange operators may have a second interest in offering content, at least in the play money case, and that is to give traders reasons to come visit the site even in the absence of new information to trade on. Content can help build trader activity.
Mike, thanks for this comment. It’s a good one.
~alex
UsableMarkets Reports from the NYC Prediction Market Conference (Yes, a little, er, a lot, late) | Midas Oracle .ORG // Oct 12, 2007 at 8:39 am
[...] Contract Content which discusses “whether having related content and links appear alongside a contract make [...]