Analyze and Forecast Team Velocity

This skill tracks team velocity and analyzes sprint performance for agile teams, calculating trends, forecasts, and actionable capacity insights. Story-point velocity counts only completed stories meeting the Definition of Done, tracked over a rolling 3-6 sprint window for stability, adjusted for team composition changes, and excluding spikes or research tasks. Capacity-based velocity instead tracks available team hours against delivered points, accounting for holidays, PTO, meetings, and non-development work, to produce an effective capacity percentage for more accurate sprint planning.

Use it when you need data-driven sprint planning and release forecasting based on historical velocity, not a single sprint's raw point count taken in isolation. It is not meant to be used without confidence intervals: recommendations should always come with a trend indicator and a confidence range, since a single average velocity number hides both trend direction and volatility.

Given sprint history, it computes average velocity and completion rate (completed versus planned points), and analyzes the last six sprints for trend direction (increasing, stable, or decreasing based on a linear fit), volatility (standard deviation), a confidence interval around the average, and a recommended planning velocity discounted by 20% of the volatility to avoid over-committing. Release forecasting adjusts the planning velocity by confidence level (subtracting half a standard deviation at 90% confidence, three-tenths at 80%) and divides remaining backlog points by that planning velocity to estimate sprints remaining. Sprint capacity planning sums each team member's available days (minus PTO) times daily development hours (minus meeting time), subtracts ceremony overhead, and applies a 15% buffer for unexpected work.

def calculate_velocity(sprints_data):
    """
    Calculate team velocity from sprint data
    sprints_data: list of dicts with 'sprint', 'completed_points', 'planned_points'
    """
    total_completed = sum(sprint['completed_points'] for sprint in sprints_data)
    average_velocity = total_completed / len(sprints_data)
    
    return {
        'average_velocity': round(average_velocity, 1),
        'total_sprints': len(sprints_data),
        'completion_rate': round((total_completed / sum(sprint['planned_points'] for sprint in sprints_data)) * 100, 1)
    }

For organizations running multiple teams, velocity normalizes across different story-point scales by computing velocity per person and a complexity factor derived from each team's average completion rate, so teams can be compared on a common footing. Predictability is tracked through velocity standard deviation, sprint goal achievement rate, scope-change impact, and a technical-debt velocity tax, feeding into a standard sprint velocity report covering completed and planned points, completion rate, the rolling 6-sprint average, trend, and team capacity.

Scrum masters and delivery leads who need velocity data to inform sprint planning and release forecasting - identifying velocity blockers (recurring retrospective impediments, story cycle time, rework and bug-fix time, external dependencies, technical debt) and applying named optimization levers: stable team composition, consistent story sizing, a clear Definition of Done, continuous retrospective-driven improvement, and automated velocity tracking.