Leaderboard statistics only include registered users.

Users' names are anonymized by default. To make your name public on these leaderboards, uncheck 'Hide your name' on your profile here.

The average absolute difference between predicted and true values, normalized by baseline standard deviation. Lower scores indicate better prediction accuracy. Each prediction is first standardized by dividing by the baseline standard deviation before calculating the error.

To be ranked in this metric, users must have made at least 10 numeric forecasts with reported results across at least 5 studies.

The average difference between a user's predictions and the mean prediction for each question. A negative value means the user's predictions were better than average. Each prediction is first standardized by the baseline standard deviation.