The standard method that forecasting platforms use to predict probabilities of future events is to average predictions of different forecasters, giving more weight to forecasters known for better accuracy. ExplicitBias is a proposed platform that will apply a similar approach to evaluating politically controversial claims.
On the platform, users will submit predictions for two types of questions:
1. Politically divisive questions that will be unambiguously resolved at some later point. Forecasters’ performance on these questions will be used to evaluate their accuracy and political bias.
2. Controversial questions that do not have a generally accepted method of resolution (e.g., “Does raising the minimum wage generally increase unemployment?”, “Does the COVID-19 virus have natural origins?”, etc.). The platform will publish averaged predictions of the forecasters, giving each individual prediction the weight corresponding to their author’s accuracy and bias.
Compared to traditional information sources (mass media, government agencies, fact-checking websites, etc.) the platform will have several significant advantages:
Impartiality: The aggregate predictions published by the platform will be independent of the political biases of the majority.
Absence of entry barriers and scalability: Participation will be open to the public and the accuracy of the platform’s predictions will rise with the increasing number of participants.
To kickstart the new platform we want to organize a forecasting tournament dedicated to resolving several major political controversies. In order to attract the maximum number of forecasters and encourage effective collaboration, we are seeking external funding to give top-performing participants awards for: (1) forecasting accuracy; (2) best supporting rationales.
As a preliminary step towards organizing the tournament, we conducted a simulation with 1160 volunteers. A brief description of the simulation’s results and their relevance to the main project can be found here:
https://www.explicitbias.com/sim_summary.html
We are seeking a total budget of $4750 to fund the prizes in the forecasting tournament:
· 1st place: $1,000
· 2nd place: $500
· 3rd place: $250
Additionally, 30 prizes of $100 each will be awarded for the best supporting rationales to incentivize information sharing between participants.