群體智慧預測網站

sec2100

https://www.almanis.com/

sec2100

經濟學人對該網站的相關報導:

https://www.economist.com/news/b ... groupthink-new-type

sec2100

Are crowds full of wisdom or irrational mobs? The answer seems to depend on how their estimates are compiled. If their guesses are generated independently (the number of jelly beans in a jar, the weight of a cow at the country fair), then the crowd can be pretty accurate. But things can change if people are aware of others’ opinions. There is a tendency towards “anchoring” (basing one’s estimate on a fixed number, such as the average guess) and overshooting (the process that creates bubbles and crashes in financial markets).

sec2100

A new website, Almanis, takes a different approach. Those who join are given an allocation of 1,000 points which they can stake to back their forecasts on the probability of political, business and economic events occurring. Predictions are not merely a matter of saying Yes or No. The more points forecasters stake, the further they can move the “price” of an event, meaning the collective reckoning of its probability. This may reduce the tendency for anchoring in forecasts, since participants will be rewarded if they successfully go against the consensus.