Check the forecast
Use the recorded outcomes to evaluate each forecast under the stated measure.
Continue your unfinished round
Your answers are saved in this tab. Continue from where you left off.
Continuing uses no additional practice start.
Discard the unfinished round?
Your answers from this round will be discarded. You can choose your settings before starting a new round.
Starting the new round uses one practice start.
Distinguish calibration from separation between forecast bins.
Your last settings are ready. Change them whenever you like.
6 questions · No time limit
Example
Example
Example · Standard
Both forecasts match the frequencies within their own forecast bins. Which separates groups with different observed event rates more?
| Group | Events observed | Forecast A | Forecast B |
|---|---|---|---|
| C | 9 of 20 45% | 55% | 35% |
| B | 5 of 20 25% | 15% | 35% |
| A | 1 of 20 5% | 15% | 35% |
| D | 13 of 20 65% | 55% | 35% |
A forecast bin contains all events given the same probability. More variation between observed bin rates means more separation, weighted by bin size.
These are fictional, completed events. The comparison describes this batch; it does not establish calibration or accuracy on future events.
- Questions answered
- 0 / 6
- Active time
- 0:00
Ready to start?
Compare forecasts A and B using the measure in the question. They may tie.
This practice could not continue. Reload the page to try again.
Reload pageOptional and off by default. Share starts, completions, replays and error categories. For visits from ChatGPT, starts and completions can also include the first QuickenBrain page and broad country. No answers, drawings or names.
How this data is usedHelp
Calibration, separation and squared error describe different aspects of a forecast. Repeated predictions share one forecast bin. The squared-error score rewards accurate probabilities on average, not certainty for its own sake. Results from one finite batch do not establish future performance. These are original reasoning exercises, not a validated intelligence assessment.
Each round has six questions and no time limit. Choosing an answer locks it. Review the steps, then move to the next question. Finish all six to save the round. Active time includes reviewing the steps, but does not count time when this page is hidden. Speed does not affect your score.
Use Tab and Enter or Space to choose an answer. The number keys also select answers, except when another control is focused. After an answer, Enter goes to the next question.
Original practice puzzles, not a standardized assessment.
About this practice
Learn a method
Compare forecasts with outcomesUse the recorded outcomes to evaluate each forecast under the stated measure.
Report a problem
Optional and off by default. Share starts, completions, replays and error categories. For visits from ChatGPT, starts and completions can also include the first QuickenBrain page and broad country. No answers, drawings or names.
How this data is used