Time series and forecasting · insight-0203
Widen or label scenario forecasts because they exclude uncertainty in the assumed drivers
- Evidence type
- documented fact
- Difficulty
- medium
- Roles
- Data scientist, Data analyst
- Industries
- aviation, energy, finance
Summary: situation, action, reason, boundary
Situation
Finance asks for next quarter's demand under three fuel-price scenarios, each with an 80% interval.
Action
Label scenario intervals as conditional on the assumed driver path. Show the spread across scenarios next to the within-scenario interval, and add a driver-uncertainty band when you can forecast the driver.
Reason
Scenario-based prediction intervals do not include the uncertainty in the future values of the predictors, so they understate total uncertainty.
Boundary
Planning, budgeting, stress tests, capacity what-ifs with assumed macro or price paths.
Symptom or risk
Each scenario interval looks narrow and decision makers read it as the total uncertainty, although the driver itself is uncertain.
Exceptions
When the driver is a decision you control, conditional intervals are the right object.
Trade-offs
Adding driver uncertainty makes ranges wider and harder to sell, but it prevents false precision.
Pitfalls
Averaging the three scenarios into one 'expected' forecast without probabilities assigned to them.
Example
Hypothetical: scenario intervals of plus or minus 4% each, while the scenarios themselves differ by 12%.
Lessons and concepts
Claims and their sources
Prediction intervals for scenario-based forecasts do not include the uncertainty associated with the future values of the predictors.