Sprint velocity vs deadline calculator

Sprint deadlines — the math your PM should run before promising a date

The industry standard is to commit to dates based on average velocity. That's why most software projects ship late: average velocity assumes a 50% confidence level. You'll only hit that date about half the time.

The right formula

adjusted velocity = avg velocity − (z × std deviation)
sprints needed = (backlog × (1 + scope creep)) ÷ adjusted velocity

Where z is the one-tailed normal-distribution z-score for your target confidence:

• 50% — z=0 (no buffer; coin-flip)
• 70% — z=0.52
• 85% — z=1.04 (recommended for external commitments)
• 95% — z=1.65 (FAANG-scale conservative)

Why scope creep matters more than velocity variance

Surveys of post-mortem'd software projects (Standish Group's CHAOS report, Digital.ai's State of Agile) consistently show 15-25% scope growth between kickoff and ship for the median project. If you don't bake in scope creep, your math is wrong before sprint 1 ends.

The two highest-leverage moves to hit a date:

1. Cut scope — recommended scope cut output tells you exactly what % to defer
2. Reduce variance, not raise mean — a team with velocity 32±2 will hit dates more reliably than 35±10

The standard mistakes

• Using best-3-sprints velocity — only the most recent 6+ sprints, including the bad ones, predict the next 6.
• Ignoring holidays / on-call rotations — derate velocity by ~10% for the holiday quarter.
• Counting points that haven't been refined — unrefined points are essentially infinite. Estimate first, plan after.
• Promising the deadline based on average velocity — that's a 50% confidence commitment dressed up as a plan.