It, somewhat densely, goes: "The squared error of the collective prediction equals the average squared error minus the predictive diversity".
But another reason you may see a lot of "honey moon," or similar shades, over the next few months, is because collective prediction processes become self-fulfilling prophecies if enough companies act upon them.
It represents, alas, our best collective prediction of what the future will bring.
To the extent that market prices can be interpreted as collective forecast, prediction markets disseminate, or broadcast, information.
Such studies have for example investigated the minimal requirements for fads to spread [8], for groups of people to make correct collective decisions or predictions [9], and the conditions for a diversity of opinions (as opposed to a widespread consensus) to be the result of communication [10].
Economists have also tried to marshal collective wisdom with "prediction markets," where people's bets on all sorts of things are distilled into a remarkably accurate forecast.
There are plenty of other sites that also try to generate collective wisdom using prediction markets, including HubDub, Trendio, and Yahoo's Tech Buzz Game.
Accordingly, it yields bad predictions at the collective scale: the predicted first follower's mean latency is far too low while the mean duration increases continuously with an increasing group size (Figure 3B C).
In [ 50], in terms of classification performance, the authors compared prediction by collective likelihoods (PCLs), based on the concept of EPs, with other classification algorithms, including decision trees, SVMs, and k-NNs.
