Augur Dispatch

Forecast accountability

Calibration, when the sample earns it.

Calibration asks whether Augur's stated probabilities match what actually happened. A 70% call should resolve as a hit about seven times in ten over a large enough sample. These numbers use only probabilities written when calls were published and outcomes finalized by an operator.

Calibration is locked

The page unlocks after 25 graded calls with an author-time stated probability. The current qualifying count is 0 of 25.

No curve or substitute score is shown before that gate. Older calls without a stated probability are not reinterpreted, and grader confidence is never used as forecast confidence.

Keep exploring

Move on to Augur's other sections.