Why a rolling average, not last sprint's number
Last sprint's velocity is one data point - it might be high because nothing interrupted the team, or low because two people were out sick. A rolling average of the last four or more sprints smooths that variance into a number worth planning against. This calculator does that math and shows the spread around it, not just the average.
How much history you actually need
Fewer than four sprints and you're mostly averaging noise, not measuring a real pattern. With four or more, real signals start to show: whether velocity is stable, trending up as the team gels, or trending down - which is worth investigating rather than quietly adopting as the new normal, especially after a drop of more than 20% across two sprints in a row.
Plan against the range, not the average
A single decimal average implies more precision than sprint-to-sprint variance actually supports. The range this calculator shows - average ± one standard deviation - is the honest planning input: it tells you what your team's capacity has actually looked like recently, not a number that discards how much it swings.
What velocity can't tell you
Velocity is a capacity signal, not a quality one. It can't predict work that hits unexpected complexity mid-sprint, and it says nothing about whether the underlying estimates are well-calibrated. Full context on using this well: using historical velocity for sprint planning, and why velocity alone is a poor productivity metric.