Estimate how long a new project will take based on historical team performance.

Online demo: https://skillfire.co/project-forecaster

It gives an objective view of the uncertainty around project estimation based on historical data.

Despite having the word estimation in its title, this qualifies as a #NoEstimates technique!

Single file version is monte-carlo-estimation.html using Ractive.js, ported from the older meteor.js version.

Enter the number of user stories completed by a team over a number of sprints, e.g.

In the box Completed stories in past sprints you would enter

5 6 5 7 6 9 12 3 4 8

The other input is the number of stories in an upcoming project, say 85. The tool is limited to between 1 and 1000 stories.

Then press the Simulate 10,000 times button and the site will run 10,000 simulations and calculate a distribution reflecting the range and frequency of project durations.

• You could alternatively use story points (or any other units) instead of number of stories, which would make this a #Estimates technique. ;-)
• Sprint length is similarly arbitrary.
• Additional stories that occur due to scope expansion, and time diverted to other areas by the team is not reflected. You'll need to allow for that yourself.

• Ractive.js or Meteor
• Bootstrap
• Google charts API

I was introduced to Monte Carlo estimation by Adrian Fittolani. He has a blog post with spreadsheet implementation and video documenting his approach. My method is simpler (no use of Takt time or re-sampling).

My frequent collaborator Tim Newbold got this project started and kicked off the coding.

CreativeCommons Attribution-ShareAlike-4.0