Risk and uncertainty modelling shows how a decision can perform across several plausible conditions. A single central estimate conceals the range of outcomes, the links between risks and the combinations that could cause serious loss. Marketways represents those ranges, dependencies and consequences, then tests which assumptions change the decision. Leaders can see where protection is worth its cost and where better evidence would reduce uncertainty.
The decision this method supports
We use Risk & Uncertainty Modelling to help clients answer: Which uncertainties could materially change the result or decision?
How the method works
Risk modelling represents adverse events, exposure and consequences. Uncertainty modelling makes visible what is not known about the inputs, relationships or future conditions. The result is usually a range or distribution of outcomes rather than one supposedly precise estimate.
A business example
An investment may appear profitable under the central assumptions. A model can vary demand, price, cost and timing together, including plausible dependencies. Leaders can then see which conditions threaten viability and where additional evidence would be most valuable.
How the client uses the result
Risk modelling shows where a decision can fail, how severe the consequence could be and which uncertainty deserves further evidence or protection. Leaders can compare downside exposure instead of accepting a single optimistic estimate.
What we deliver
We produce forecast ranges, scenarios, risk distributions, prioritised alternatives or an optimised plan. A manager should understand which assumptions drive the result and what conditions would trigger a different action.
Limits and complementary methods
A range is more honest than a point estimate, but the range is still shaped by model structure, assumed dependencies and events the analysis did not anticipate.
Selected methods and techniques
We select from these established methods according to the decision, evidence and operating conditions.
- Break-even analysis: Define the financial result that must reach zero and solve for the price, volume, utilisation, time or other driver at which that happens. For operating break-even, distinguish fixed costs, variable costs and contribution per unit; for cash or discounted break-even, use the relevant cash-flow timing and state the measure explicitly. Test multiple products, capacity limits and uncertain assumptions where they matter. Break-even is a boundary calculation, not proof of profitability, liquidity or investment value beyond that point.
- Copula modelling: Fit or specify each uncertain quantity's marginal distribution, select and estimate a copula to join them, and test whether the resulting joint model captures relevant dependence, including co-movement in the tails. Use the fitted distribution to simulate or calculate combined outcomes. Results depend on the marginals and copula, and sparse extreme data can make tail dependence highly uncertain.
- Correlation / dependence modelling: Describe how variables move or occur together and how strongly they are linked. Such a link can help predict outcomes but does not on its own show that one variable causes the other.
- Distribution fitting: Find a mathematical description of how often different values occur that reasonably matches the observations. Check both typical values and unusual values important to the decision; fitting the average alone is not enough.
- Expected-loss modelling: Estimate the average loss across possible outcomes over a stated period, using both how often events may occur and how much each may cost. Account for risks that can occur together; the average is not the most that could be lost.
- Frequency-severity modelling: Model the number of loss events over a stated period separately from the size of each loss, then combine the two to estimate aggregate loss. Represent dependence, changing exposure and conditions that affect both event count and severity. The model's annual loss distribution is different from the distribution of loss conditional on one event.
- Monte Carlo simulation: Calculate many possible outcomes by repeatedly drawing plausible values for uncertain inputs. The resulting spread shows what the specified model can produce; its usefulness depends on the input assumptions and how the inputs are connected.
- Parameter uncertainty analysis: Assess how uncertain the estimated numbers inside a model are and how much that uncertainty changes the conclusion. These estimated numbers are the model's parameters.
- Prediction intervals: Construct and evaluate lower and upper bounds intended to cover a future observed value at a stated rate under a specified forecasting model. Define the target, forecast vintage, horizon, coverage level and conditioning information; include both uncertainty in the estimated forecast and the variation expected in the future observation; and check empirical coverage and width on comparable later forecasts. The method is the procedure used to calculate and calibrate the bounds. A Prediction Interval is the populated output, a forecast interval is the wider ontology term that may require its interval type to be named, and an assumption-based scenario range is not a calibrated prediction interval.
- Predictive uncertainty analysis: Assess how uncertain future outcomes are, separating uncertainty in what the model has learned from variation that will occur in practice even with a good model.
- Probabilistic forecasting: Forecast a distribution or probabilities for clearly defined future outcomes rather than only a point estimate. State the target, horizon, conditioning information and interval meaning, then assess calibration and sharpness on comparable later cases. A wide calibrated forecast may be honest but not useful; a narrow forecast is not useful if overconfident.
- Probabilistic modelling: Represent an uncertain event or quantity through its possible values and probabilities, including dependence among related quantities. Define units, population, horizon, conditioning and the evidence or judgement behind the probabilities. The model makes uncertainty explicit; it does not remove structural uncertainty or justify probabilities for possibilities outside its representation.
Parent method family
Forecasting, Risk & Optimisation explains how this method connects to adjacent methods and relevant services.
Related service families
These service families contain business questions supported by this method. Service pages link to the wider method family so readers can understand the complete analytical approach.