Methods and settings
| Method / input | Behavior |
|---|---|
| Arithmetic mean | Must be positive; entered in the same unit as the sampled quantity. |
| Standard deviation | Nonnegative arithmetic standard deviation; zero is constant. |
Do not enter log-space parameters
The UI accepts the arithmetic mean m and standard deviation s of the positive quantity. It derives the log-space parameters internally: variance in log space is ln(1 + (s/m)²), and log-space mean is ln(m) minus half that variance. Do not pre-transform the amount with a logarithm before entering it.
Interpret mode and skew
Absolute mean 100 EURk and standard deviation 20 EURk describe a positive amount whose upper tail is longer than its lower tail. The median is generally below the arithmetic mean. In Deviation mode the sampled increments are themselves positive: this does not produce symmetric upside and downside around the current input. Use another family when negative deviations are part of the intended uncertainty.
Bounds change the resulting moments
The configured arithmetic mean and deviation describe the original distribution. Conditioning it on absolute source bounds changes the realized mean and dispersion. Check retained probability rather than assuming the bounded sample keeps the original moments. Lognormal is not offered for integer inputs.
Documents the current implemented behavior. Numerical examples are synthetic, not market assumptions.