Just to be clear, the probability weighted forecast isn't about yielding an accurate forecast number, it's about testing the business model against various scenarios to understand how it will react to those, both optimistic and pessimistic. From a methodology perspective, it's the exact opposite of trying to build a single accurate forecast.
It's acknowledging that you aren't in a position to build a single highly confident forecast, so instead you build a set of different forecasts, and assign a probability to those, for example, I'm 30% confident in the pessimistic forecast, 60% confident in the baseline forecast, and 10% confident in the optimistic forecast. We then plug those in, and see how those impact the model.
Then, you can start to test the model - ok - what happens if we push the pessimistic to 40% and baseline to 50%? Or, what if we increase the volume on the optimistic scenario. Or, you can start to look at it from another perspective, such as what are the minimum monthly sales volumes necessary to break-even? Or, at what point can I actually draw a salary, have enough cash to make a subsequent investment, what level of sales do I need to be able to afford a certain level of inventory build, etc.