TryChange Chosen probability contour from 0.1 to 0.5.
ObserveCompare First whole population above 10.0% with First whole population above 50.0%.
ExplainExplain why the 10% reference contour is not a general safety limit.
Technical boundaries. The birthday approximation assumes independent uniform choices. The selected contour requires a consequence that justifies it and a model matching the real allocator.
Choose the Probability Target Explicitly
For (n) independent uniform choices from (m) values, the birthday approximation is
P(at least one duplicate)≈1−e−n(n−1)/(2m).
Solving for the population at a chosen probability (p) gives
n≈−2mln(1−p).
For a 10% reference contour, the multiplier is about (0.459m). Ten percent is not a general safety limit; it is simply a visible contour learners can compare. An identity used for billing, safety, or actuation may need a much tighter target and coordinated prevention. A recoverable laboratory token might accept more.
{const m =2** designBits;const lambda = designPopulation * (designPopulation -1) / (2* m);const probability = lambda >700?1:1-Math.exp(-lambda);const contour =Math.ceil((1+Math.sqrt(1-8* m *Math.log(1- designTarget))) /2);returnhtml`<div style="background:#f8f9fa;border-left:4px solid #16a085;border-radius:8px;padding:1rem;font-family:system-ui,sans-serif"> <p style="margin:0 0 .6rem"><strong>Random-choice model:</strong> ${designPopulation.toLocaleString()} choices from 2<sup>${designBits}</sup> values</p> <table style="width:100%;border-collapse:collapse"> <tr><th scope="col" style="text-align:left;padding:.35rem">Model quantity</th><th scope="col" style="text-align:right;padding:.35rem">Value</th></tr> <tr><td style="padding:.35rem">Approximate chance of at least one duplicate</td><td style="padding:.35rem;text-align:right;font-weight:700">${(probability *100).toFixed(probability <0.001?5:2)}%</td></tr> <tr><td style="padding:.35rem">First whole population above ${(designTarget *100).toFixed(1)}%</td><td style="padding:.35rem;text-align:right">${contour.toLocaleString()}</td></tr> </table> <p style="margin:.7rem 0 0"><strong>Release question:</strong> does this model match the real allocator, and what consequence justifies the selected contour?</p> </div>`;}