Figures / Sources

300

Random records read before a delivery

Our setting, read from the engine (RISK_POLICY.reviewDelivery). With no identifier found in n records drawn at random, the one-sided 95% upper bound on the share of records still holding one is 1 - 0.05^(1/n); for n = 300 that is 1.0%. The method follows the upper-confidence-limit test used for EMA Policy 0070 clinical reports.

Updated 15 Oct 2026, from its newest record.