We’ve been talking a lot about the increasing importance of people to search, both directly through networks of people actively recommending & exchanging sites, and indirectly through search engines’ efforts to ape human wants.
Obviously, the same social thing has been happening with news via sites like Digg, del.icio.us and so-on, but it’s interesting to see it happening with news prediction, which is a new one on me.
Yahoo’s applied to patent a way to index and retrieve information alongside a related date. Searching for ’space’ events might produce a list of future events by year.
An example they give for 2034 would return:
A couple of things occur to me about the usefulness of the approach when it comes to forecasting near-future events. It’s fine to have 100 results for 2034, but there might be 100,000,000 results for 2014. How does that help the user?
Also, the two examples above were sound predictions in 2005. By 2030 they may be laughable - Voyager 2 might in fact have been eaten by a huge alien that takes up residence on the moon. Fear him, puny humans.
The results need to be nimble enough to respond quickly when reality fails to follow the amassed weight of earlier expectations.
Which makes me wonder if a better way to forecast near-term events is just to let people do it based on their own knowledge, prejudices and gut feeling, as Hubdub attempts. You can bet ‘play money’ on the outcome of news stories, and the most successful news-guessers are ranked accordingly.
If early promise translates to later success, expect to see socially-generated news predictions creating or informing news stories. A combination, perhaps, of the unnamed “analysts” or “commentators” beloved of the journalist, with the way that journos already use Facebook and other social networks.
VIA SEO by the Sea
IMAGE by Flickr user emdot, republished under a Creative Commons Attribution license.
February 25th, 2008 at 1:41 pm
>You can bet ‘play money’
Or real money. You’re talking about prediction markets. They’ve been used and studied for some years in political and economic forecasting. Ironically, since you mentioned user prejudices, the idea is that actually betting resources on outcomes lessens predjudices and makes people more truth-seeking - an actual implementation of the saying “put your money where your mouth is”, as it were - and this results in better forecasts.
The only name in this field I can think of off-hand is the economist Robin Hanson (google him). He advocates, among other things, that prediction markets should be used to set governmental policy. Very interesting reading.