Monday, July 30, 2007

Predicting Potter and Consoles

Two public prediction markets have recently reached their end: June game console sales and “Will Harry Potter survive the Deathly Hallows?”

What makes comparing the results of these two markets interesting is the different type of question they are predicting. In the case of the console market, they are attempting to predict the buying habits of a large group of people as well as the production capacity and marketing moves of large companies. On the other hand, the Harry Potter market is attempting to predict the creative whims of a single person (J.K. Rowling).

simExchange’s Game Console Market (Full Report)

I particularly like the results of this market because is that there is a different predictive agent used for comparison.

As you can see, the results are mixed. The market predicted PS3 sales the best, missing the actual sales by just 0.1%. On the other hand, I missed Wii sales by over 13%. While the Wii prediction is not nearly as accurate as the PS3 prediction, the market did outperform the analyst in 3 out of 5 cases. Previous months continue this trend with the market beating the analyst 4 to 1 in May and 5 to 0 in March. Only in April did the analyst out predict by a factor of 3 to 2.

Harry Potter (Warning: spoiler alert)

While there is no independent predictive agent for this market, we can compare the results of multiple public markets to gauge their accuracy. The day before the final book was released, the chance of Harry Potter surviving was at 75% on NewsFutures and 84% on Inkling Markets. For the lifetime of both markets he was predicted to survive. Although this turned out to be an accurate prediction in this case, this type of question is not answered as reliably as the game console market. In a recent New Yorker article, James Surowiecki, author of the Wisdom of Crowds, wrote that questions that require predicting how large groups of people will behave are much more successful than predicting a the actions of a small number of people. In the case of the ending to Harry Potter, they are attempting to predict how just one person will behave.

Although guessing what will happen to Harry Potter is a much simpler task than trying to predict what will happen with entire console market, the predictions of the console market will most likely be more accurate over time than markets that attempt to predict book endings.

Found in (0) Comments • Permalink http://www.sundog.net/index.php/sunblog/entry/predicting-potter-and-consoles/

Like this post? Subscribe to our RSS feed.