Remove ANOVA fields from Convergence Average Loss Table (CALT)
The standard error in the Convergence Average Loss Table (CALT) has been observed to overestimate the observed sampling error. This is because the random effects model used to partition the variance into vulnerability and hazard factors requires those contributions to be random. However, the hazard element in the Oasis framework is fixed, not random: events occurrences are assigned to years in a fixed timeline. Therefore, the hazard element in the variance of the Average Annual Loss (AAL) does not reduce with increasing samples, leading to a larger standard error.
As the ANOVA (ANalysis Of VAriance) fields are not helpful in predicting AAL convergence, they have been dropped. The standard error is calculated as s / sqrt(IM), where s is the standard deviation of the annual losses, I is the total number of periods and M is the number of samples.
Remove ANOVA fields from Convergence Average Loss Table (CALT)
The standard error in the Convergence Average Loss Table (CALT) has been observed to overestimate the observed sampling error. This is because the random effects model used to partition the variance into vulnerability and hazard factors requires those contributions to be random. However, the hazard element in the Oasis framework is fixed, not random: events occurrences are assigned to years in a fixed timeline. Therefore, the hazard element in the variance of the Average Annual Loss (AAL) does not reduce with increasing samples, leading to a larger standard error.
As the ANOVA (ANalysis Of VAriance) fields are not helpful in predicting AAL convergence, they have been dropped. The standard error is calculated as s / sqrt(IM), where s is the standard deviation of the annual losses, I is the total number of periods and M is the number of samples.