Have you ever made an online prediction about who will win the next Champions League? Did you ever take a look at a tennis player's past winnings so you could predict who will be the next ATP leader? Then you have already contributed to the crowdsourced forecasting community.
Crowdsourced forecasting is the process of a community making future predictions. The community analyses present and past data or other sources of information. Sometimes, forecasters appeal to their own general knowledge and personal experiences.
The Brier score is a function that measures the accuracy of probabilistic predictions. It's used when predictions must assign different probabilities to mutually exclusive results. All probabilities must sum to 1, since each individual probability is in the range of 0 to 1. The Brier score is a tool of calibrating probabilistic predictions. It is also known as a “cost function”.
Our forecasters make predictions on a series of events. These comprise of 4 particular categories: economics, politics, entertainment and sports. On the right, you have a display of the evolution of the Brier score. Each category shows the accuracy of all-time predictions.
Check out a series of events predicted by our forecasters!
Each potential outcome has a different likelihood of coming true. That's because there are different degrees of probability given by forecasters.
So how did the Brier score get to be so accurate?
The community grew from a few predictors to a large panel of knowledgeable forecasters. The general Brier score is ChecktheCrowd’s statement over time. It shows the accuracy of forecasts blended into a unique recipe of data analysis and wisdom of the crowds.
The unique recipe put into our product consists of a mélange of data analysis combined with our community’s predictions. It can predict specific results in sports, entertainment, economics and politics. Our inspiration stems from the ox parabola in which hundreds of people had to guess the weight of an ox. Some of them estimated too high while others guessed too low. Unexpectedly, the average came very close to the actual weight. This proved that the all-time accuracy of the crowd was impeccable, even though many individual guesses were incorrect. We present you: CheckTheCrowd.