In May 2013 at Drupal Camp Dublin I gave a presentation titled "The lost art of estimation" it outlined some approaches and methods which I hoped would aid in everyone getting better, more accurate estimates. You could be forgiven for thinking that the naming of this session means I've changed my mind about estimates but, I still believe they're valid... in the right situation. The trouble tends to be with how estimates are used and as the figures we produce will always be used in a particular way, we should make sure to provide sufficient supporting information whenever figures are in the wild.

In this session, I'll outline the issues with using estimates, averages and other single-figure predictions and then walk-through an alternative approach using probabilities and Monte Carlo Simulations.

You do not need to be a mathematician to attend this session, everyone will come away with an understanding of:

How estimates can give a false picture for a project.
The cognitive bias involved during an estimation process.
What data, at minimum, should be captured.
When to run forecasts.
How to communicate probabilities.
Whilst this session will show the benefits and potential of forecasting, a subsequent BoF/Workshop has been proposed to allow everyone to be guided through their first use of the available Open Source tools.

https://events.drupal.org/vienna2017/bofs/forecasting-tools-project-teams