Sequential policy-risk update

Act when evidence earns it.

Turn dated observations into a transparent posterior path. See which update changes the decision state without pretending uncertainty disappeared.

PRIOR PROBABILITY

ACTION THRESHOLD

LR above 1 confirms; below 1 disconfirms. Weight 0–1 discounts evidence. Calibrate inputs externally and avoid double-counting correlated sources.

Load the sample or enter at least one evidence row.

—observations

—final posterior

—first crossing

—final update

Posterior path

Update trace

—posterior odds = prior odds × LR^weight

Prior sensitivity

—Final posterior using priors 10%, 20%, and 30% with the same evidence.