Project Forecast

In software development, accurately predicting project timelines is a challenging yet essential task.

Project Forecast leverages the power of Monte Carlo simulation to provide data-driven predictions.

By analyzing historical data, this method generates realistic timelines, helping teams manage risks, allocate resources, and hit deadlines more consistently. In this article, we’ll explore how LinearB’s Monte Carlo simulation works, its benefits for project management, and how it can elevate your forecasting accuracy.

Getting Started

To access Project Forecast, navigate to Projects and then click Forecasting. Once the view loads, click the Forecast tab. You should see a screen that looks like this:

If you haven’t created any projects yet, start by clicking Add Filter. This opens the filter selection component, where you can choose Jira projects, Initiatives, epics, or Azure Boards Epics, Features, tags, and more to define your project’s scope. For more details, see Saving Filter Sets and Projects.

To ensure the best results, make sure the following conditions are met:

• The selected scope contains enough issues to be meaningful (typically over 30 issues).

• There’s a recent history of issues moving to “Done” within the past 5 weeks.

• The scope is not continuously expanding. In many cases, selecting an entire board or project may represent ongoing work, which doesn’t have a defined endpoint.

Generally, the best approach is to select a set of active Jira Epics or Azure Boards Features that are currently being worked on.

Understanding Results

Prediction analysis is displayed as a color-coded chart with three distinct views.

In the default view, you’ll see a weekly timeline for the upcoming weeks, overlaid with colored bars indicating the probability of project completion each week. The taller the column, the higher the probability of completion in that specific week. A red bar indicates up to a 50% chance of completion, signaling a low likelihood of finishing on time. In contrast, deep green represents a 95% or higher probability, indicating strong confidence in timely completion. Other colors in between correspond to varying probabilities of success.

The second view will be familiar to those who have used Monte Carlo simulations, as it uses the typical chart format for presenting results. This chart shows a standard distribution (or bell curve), with the tallest columns representing the weeks with the most frequent completion estimates. Color coding ranges from red (low probability) to deep green (high probability), as described above. Additionally, markers are overlaid to indicate delivery probabilities between 50% and 95%.

The third, more intuitive view is tailored for those less familiar with Monte Carlo simulations, featuring a calendar where upcoming weeks are color-coded from red to green, showing delivery probability. You can navigate through months, view today’s date, and see the target completion date and its probability of success

Errors and empty states

In some cases, forecasting may not produce meaningful results. When this happens, the system presents an empty state with explanations and guidance on how to resolve the issue.

• If the selected filter set or project has few or no tickets from the past 6 weeks, we’ll consider it empty and display the following message

  

• If the model predicts less than a 3% chance of completion within the next 15 weeks, we’ll display a ‘Forecasting Unavailable’ message

  

• If 95% of the issues or work items are marked as ‘Done,’ we’ll display a ‘Great Job!’ screen to celebrate project completion