Methods and settings
| Method / input | Behavior |
|---|---|
| Normal | Symmetric uncertainty using mean and standard deviation. |
| Lognormal | Positive, right-skewed outcomes using arithmetic mean and standard deviation. |
| Uniform / Triangular / PERT | Bounded uncertainty with equal likelihood or a most likely value. |
| Discrete / Empirical | Explicit weighted outcomes or resampled observations. |
Start with the evidence
Use a bounded distribution when you can defend minimum and maximum values. Use a most-likely value only when it expresses a real modeling judgment. Use Discrete for a small set of explicit outcomes and Empirical for a list whose observed frequencies you want to retain. Normal and Lognormal require meaningful dispersion assumptions; a standard deviation is not a low/high range.
Check mode, unit and support
The same parameters mean different economic assumptions in Absolute and Deviation mode. Check the current input, unit, allowed source range and optional absolute bounds before saving. A mathematically valid distribution can still generate invalid financial outputs, such as a valuation case without a meaningful denominator. Start with a small test and inspect validity before increasing run count.
Parameters are assumptions, not fitted estimates
BalanceCheat does not select or fit a distribution from company history automatically. Empirical entry is not an uploaded time-series forecasting model. The family pages explain exact parameter semantics and useful synthetic examples. All examples illustrate configuration; none are suggested market assumptions.
Documents the current implemented behavior. Numerical examples are synthetic, not market assumptions.