I made a game that teaches beginner calibration. Have you ever wanted your brother/girlfriend/aunt/dog to make more calibrated probabilistic claims? Now you can send them Calibrate! (Or you can play it yourself.)
Semi-big catch: Calibrate was mostly made for computer science students, so non-computer science students may have a hard time getting very high scores. People from other hard sciences have also enjoyed it, though!
More on Calibrate
The premise of the game is to get rich by investing in a calibrated way. It's more fun than existing calibration training tools in my opinion, because there's more sense of progress from earning money and moving through the game. It's not very long, roughly 30 minutes total.
I also did a very low-powered analysis of calibration outcomes with 11 play testers. The results look promising; see the calibration curve on the left. See the plot on the right for how to interpret the curve. On a cursory look, outcomes are looking similar to those of other calibration trainings, which are usually much longer. (Confidence interval calibration didn't work super well compared to multiple-choice confidence calibration; see the unhappy green X on the left.)
I made a game that teaches beginner calibration. Have you ever wanted your brother/girlfriend/aunt/dog to make more calibrated probabilistic claims? Now you can send them Calibrate! (Or you can play it yourself.)
Play now (desktop only)
Semi-big catch: Calibrate was mostly made for computer science students, so non-computer science students may have a hard time getting very high scores. People from other hard sciences have also enjoyed it, though!
More on Calibrate
The premise of the game is to get rich by investing in a calibrated way. It's more fun than existing calibration training tools in my opinion, because there's more sense of progress from earning money and moving through the game. It's not very long, roughly 30 minutes total.
I also did a very low-powered analysis of calibration outcomes with 11 play testers. The results look promising; see the calibration curve on the left. See the plot on the right for how to interpret the curve. On a cursory look, outcomes are looking similar to those of other calibration trainings, which are usually much longer. (Confidence interval calibration didn't work super well compared to multiple-choice confidence calibration; see the unhappy green X on the left.)
Play Calibrate now
Other excellent calibration tools (not computer games):
Calibrate your Judgment by Clearer Thinking
Estimation Game by Sage