2. Simple models - Combine uncertainty with simple models (at first)
Why it matters
When dealing with a problem riddled with uncertainty, it’s all too easy to get sidetracked by the details. When we focus too much on the details, we often have a hard time seeing the wood for the trees. We lose track of understanding and grappling the real problem at hand and we end up wasting time.
To get to grips with the said problem with uncertainty, a simple model is what you need. This is how it works, you make a hypothesis about what the cause of the uncertainty is. Take the most obvious thing that comes to mind.
Then test your hypothesis. Does it explain the uncertainty well enough? Can it be easily disproved? Does it rely on assumptions that are unrealistic? Just keep going and rinse and repeat until you find a hypothesis that is robust enough. The beauty of using a simple model is that it is a quick way to detect which input is causing most of the uncertainty.
How it works
See if a simple model would have forecast something that transpired recently. Test the model against something in the real world.
If not, what is the biggest reason why that comes to mind? Add this to your simple model. Rinse and repeat.