Simple Release Planning with Historical Velocity Data
Lets assume that you have 26+ past iterations to calculate velocity and there are 5 iterations left to finish product backlog. So how many backlog items can you do until the end of the project?
Velocity Calculation Steps
- As per the table below you should throw out 8 best (highest) and 8 worst (lowest) velocitiy values
We throw away outlying values because those values are unlikely to be the team’s future average. The range that remains after throwing away some number of outliers represents a confidence interval around the team’s future velocity.
- And then you can determine true velocity in the range between 36 and 41 with %90 level of confidence as follow.
This method is based on an statistical technic known as “Confidence Interval Around the Median” using a binomial distribution.
So in this case you can make the product backlog planning as follow