On this page
Read what the inputs actually provide
The tool takes a supplied same-horizon loss distribution. Its probabilities and loss values do not describe how separate periods depend on each other, how positions change or how losses accumulate. Selecting a longer reporting horizon therefore needs additional assumptions.
Distinguish a dispersion rule from a tail model
Under particular independent/increment assumptions, a variance calculation can lead to square-root scaling of standard deviation. That does not automatically establish how a finite loss quantile, a mean component or nonlinear exposures scale across periods.
Collect the missing horizon information
| Field | Question |
|---|---|
| Dependence | How do period outcomes relate? |
| Aggregation | Additive losses, compounded returns or repricing? |
| Exposure | Static holdings or changing positions? |
| Distribution | Which multi-period loss law? |
| Measure | Which confidence and quantile rule? |
Keep the tool within its scope
Construct or obtain the appropriate multi-period loss record, label its source and then calculate its summaries. The public tool performs no horizon scaling and supplies no calibrated market distribution.