In this video, PST Thomas Schissler and Glaudia Califano explain Monte Carlo Simulation. Monte Carlo Simulations can be used to make probabilistic forecasts. This video explains how it can be used in the context of product delivery.
I guess even if you are using story points, you should be able to quite easily come up with the number of items completed in a period (e.g. a Sprint). With that data you can run the Monte Carlo simulation as explained in this video. Maybe your line of thinking is that considering the size of items in the simulation might increase its accuracy, but that is not certain. My recommendation is to do this calculation with "right-sized" items. This means that you break down too big items into smaller ones that have a size you feel comfortable with. If your historic data and the items you are applying the forecast to are right-sized, the simulation would create results that might be as accurate or even more accurate as doing it based on size. Considering size adds a lot of complexity which can often be misleading. Let me know if that answers your question or if you would like to know more.
Story points are NOT a means for forecasting. Story points are simply estimations, that should largely be thrown out past the discussion they may bring to the refinement. I would recommend looking up some of the videos daniel vacanti has on predictability (Drunk Agile is a podcast he does that he goes in depth on these topics). Additionally has two books related to forecasting.
I am sorry, this might be a dummy question or I have missed out something. Because I am wondering why for the first histogram the 80% is on the left of 50% to predict how many items will be delivered in 10 iterations and in the second histogram the 80% is on the right (which for me is what I am familiar with). Thus, what have I missed out on in the first histogram?. Can anyone help here? Thanks
In the first question, the x-axis scale represents something (the number of items completed by a particular date) that is less and less certain to happen the further to the right we go. In the second question, it is the other way around, the x-axis scale represents something (date to complete all items) which is more and more certain the further to the right. So the x-axis scale in terms of probability in the two examples are in opposite directions and is why the counting is done from opposite sides. Does that help?
