What sort of forecasting is better?
- If this, then …
- It will be …
The first makes assumptions explicit. The second hides the assumptions and takes them for granted.
In my work as a strategy consultant, I make extensive use of scenarios, and often we might not consider the likelihood of scenarios while constructing them. It matters because it helps us to immerse ourselves into a reality such that our construction of the reality is not affected by how likely we think it would be. It is more important to be able to extensively work our the implications of our assumptions at that stage.
Only after the scenario modelling is complete, it makes sense to step back and examine the assumptions, perform sensitivity analysis and consider how the outcomes are sensitive to some of the assumptions.
And then what? Then we consider likelihoods of those assumptions manifesting.