Agile Sprint Forecast — Monte Carlo Simulator

Backlog & Team

Backlog size 120 pts

Mean velocity / sprint 15 pts

Velocity variability 30%

Target sprint 12

Forecast (live)

Trials run

P(finish ≤ target)

—

P50 forecast

— sprints

P85 forecast

— sprints

How it works

Real sprint velocity is noisy — scope changes, illness, spikes and interruptions mean a team almost never delivers the exact same number of points every sprint. Monte Carlo forecasting embraces that noise instead of pretending it away: rather than dividing backlog by "average velocity" once, it replays the project thousands of times with randomized velocity and reads off the spread of outcomes.

for each of N trials:
  done = 0, sprint = 0
  while done < backlog and sprint < cap:
    v ~ Normal(mean, mean·variability)   // clamped ≥ 0.1·mean
    done += v ;  sprint += 1
  record sprint as this trial's completion time

P50 = median of all completion times
P85 = 85th percentile  (85% of trials finished by here)
P(finish ≤ target) = fraction of trials with sprint ≤ target

Backlog size — total story points remaining to ship.
Mean velocity — the team's average points delivered per sprint.
Velocity variability — standard deviation as a % of the mean; higher means less predictable sprints and a wider, flatter histogram.
Target sprint — the sprint you are asked to commit to; the vertical marker and the "P(finish ≤ target)" readout tell you how much confidence that commitment actually has.

This is the same technique (throughput/velocity Monte Carlo forecasting) used by real Agile teams and tools like Jira's forecasting add-ons to answer "will we make the date?" with a probability instead of a single optimistic number — P85 is a common industry convention for a commitment date because it means only 1 in ~7 simulated futures misses it.

Finish ≤ target sprint Finish > target sprint Did not finish (>30 sprints)

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.