FAILURE MAP
← Case archive

FA-63010 / Insurance premium rating / Member archive

Loss development and trend to rating period: Remainder months are trended as whole years · case 05

Trend factors explode when the span is not a whole number of years.

Member previewVariant 5 · 3 implementations · 9 checks per implementation

Case contract

Input losses, age (months, positive multiple of 12 else 'invalid'), ldfs (age-to-age *1000, first is 12-24), tail *1000, trend (annual per mille), exp_start and pol_start (months). CDF = tail * product of ldfs from index age/12-1. Trend span = (pol_start+12) - (exp_start+6) months; factor = (1+t)^years * (1+t*rem/12). Return floor(losses*CDF*factor+1/2).

Why this case matters

Projected ultimate trended losses drive the indicated rate level for the prospective policy period.

One recorded failure

Sample boundary fixture

This sample comes from the broken implementation of a controlled reproducer.

Boundary fixtureActualExpectedOutcome
second evaluation2106041914668335Failed

MEMBER ARCHIVE

The complete case is available to members.

This record includes three runnable implementations, regression fixtures, execution results, and source hashes.

Member access is invitation-based. Sign in with your invited account to inspect the sources.

Sign in to the archive ↗