IEEE Future Directions
Related Articles
Predicting the future is probably one of the characteristics of the human species. For as far as we can go back in time we have evidence of soothsayers …
Forecasting has become a science in the last 100 years, particularly in some areas, like weather forecasting where the massive processing capacity and pervasive sensors network are producing more and more accurate forecasts.
There have been studies in many areas aiming at discovering some patterns that can be used to forecast evolution. A most famous one is the Moore’s law that has proved amazingly accurate over the last fifty years.
Other kinds of studies explored the common wisdom of a crowd and are showing that, indeed, a sufficiently large crowd can provide quite accurate forecasting that in certain areas can best the ones voiced by specialists.
A well know example is the estimate of the number of marbles (or nails) contained in a large bowl. It has been demonstrated that by asking a sufficiently large group of people (hundred or more) and averaging the answers you get to a pretty good estimate. This has become to be known as the "crowd wisdom".
Now the George Mason University has launched a tool to harvest this crowdsourcingwisdom, SciCast.
The tool supports people, not necessarily experts in specific fields, although once one register has to declare his/her fields of interest, to insert their forecast on a variety of technology, markets and so on.
When you log in (first you need to register, I did it and it is a breeze) you’ll see 14 areas and by selecting one you are taken to a list of questions upon which you may express your forecast. You can also propose a question and wait for the “crowd” to respond.
SciCast does not use a simple “averaging” but bases its forecasting on a variety of inter-related algorithms leading to a combinatorial market forecasting. This means that forecasts in an area may influence forecasts in another related area.
The system has been tested for 2 years involving 1,000 experts and the results have shown a 40% better forecasting than a simple averaging.
Experts operate in a cooperative and competitive framework. If your predictions turn out to be better than average your leadership index increases. Experts are invited to go back over and over to refine their predictions as more information become available.
At the EIT ICT LABS we have a community that spans thousands of researchers, marketeers and business people. Probably we should be able to leverage on this community also using tools like this one…
I guess this is just “yet another” example of how an interconnected Society can exploit the intelligence of its members to create an emerging global intelligence. Watch out: it is on a very small scale, so far, but the principle is similar to the one that creates intelligence out of the local -limited- intelligence present in each of the neurones in our brain!
