One of challenges of sprint planning is settling on a good choice of velocity. One simple, but imprecise and approach uses burn-up charts. A clear explanation of dealing with uncertainty of velocity using burn-up charts can be found late in  this  video. This technique may not be good enough, especially in the days of the project or if the project never settles in with a nearly constant velocity. Here is an explanation of applying some math to get a better answer if you need it.

Those who have followed me through the years may know that I have been an advocate of using Bayesian reasoning for development. The key thought is that interesting development problems deal with uncertain quantities such as team velocity or time to completion, and that Bayesian analysis is the way to reason about uncertain quantities. This last statement is contraverial is some circles (for a history click here). In any case, without getting too deep in the discussions, Bayes is a very practical answer to many problems that arise in development.

I have built an example using the AgenaRisk Bayesian Net tool (A free version can be found here.). The example is described in this document, Bayes for Velocity. The AgenaRisk file used in the document can be found here. You can download the free version of AgenaRisk to play with the model yourself. The model is based on a parameter learning example file delivered with the tool.

I hope to write a comparison of this method with the more simplistic methods in a future blog.

As always, comments welcome!