01 / Arrival distribution
When might a CRQC arrive?
Chart. Use the left and right arrow keys to move the year; Escape hides the reading.
Chart data
02 / Exposure window
The Mosca timeline.
Exposure occurs when arrival precedes the end of data sensitivity: Z ≤ s + x.
03 / Compare data classes
Longer-lived data carries more exposure.
04 / Model risk
The family matters beyond the fitted points.
Weibull
Lognormal
Log-logistic
Gold tone = scenario · right = min–max
Exact values and family key
W: Weibull · LN: Lognormal · LL: Log-logistic. Values are probabilities on a 0–1 scale.
Where the curve families diverge
05 / Monte Carlo
Allow the migration date to slip.
Shelf-life uncertainty
P(exposure), horizon view
Exposed record-years · share of draws
Every data class, same draws
06 / Sensitivity
Which assumptions move the result?
Probability scale
Input ranges used
07 / Survey weight
What changes when the signal gets a weight?
08 / Last-cohort view and HNDL cover
A liability that emerges after the harvest.
Last-cohort view
probability, arrival after migration only
09 / Method and limitations
A reproducible model, with explicit limits.
The full write-up Memo (PDF) Notebook