Monte Carlo Forecasting

Agile Project comes with a powerful Monte Carlo simulation engine. It enables delivery date forecasting showing the expected delivery date range with 50% to 85% probability. Simulation on both issue count and story points is supported.

How does it work?

The solution builds an array with done issues or story points based in the day the issues was moved to done (If an issue is done can be configured by either using the issues status or issue resolution field in settings). The simulation engines then randomly picks days in the array and by that simulates how many days are needed to match the simulation scope. It does this 5000 times and then plots the output on a distribution chart. Depending on the days picked the different simulations finish earlier or later and this gives the 50% to 85% probability delivery date.

Why Monte Carlo forecasting?

Software projects are not deterministic. This means that if you run it multiple times the delivery date will not be the same. This is driven by a large amount of uncertainties within software development such as new scope, people leaving, misunderstandings etc…

Monte Carlo simulations enables people to talk about a delivery range with probability e.g. October 2021 instead of a specific date. Like in weather forecasting software delivery dates can swing and the key thing is to do continuous forecasting. It enables you to have a richer and more content driven stakeholder discussion with early warning if things are getting wrong instead of just being late the last week!

More Information Monte Carlo forecasting

There exist a large amount of articles on Monte Carlo forecasting in software development. Please find one here as well as the Wikipedia page.