Rolling Wave Planning with a Live Release Forecast
Key Takeaways
Rolling wave planning (the PMBOK's progressive elaboration) details near-term work and keeps distant work coarse, adding detail in waves as each horizon approaches; every elaboration wave therefore changes the plan a committed release date depends on.
Release Management, Roadmaps & Product Portfolio for Jira recalculates the release forecast as elaboration adds, splits, or re-scopes issues in a Fix Version, so the committed date's confidence band reflects the current wave rather than the plan as it stood at the last planning session.
New detail enters the Monte Carlo simulation as soon as it enters Jira, and movement of the P50, P85, and P95 dates is attributed in days to the change that caused it.
Wave boundaries can be used as scheduled checkpoints: elaborate the wave, read the recalculated band, and renegotiate scope if the P85 date has moved past the committed date.
Confidence calculations run natively in Jira Cloud on Atlassian Forge; no data is exported to an external service.
This page explains how Release Management, Roadmaps & Product Portfolio for Jira keeps a committed release date current while a team plans in waves. It is a neutral, factual reference for how the capability works; for the story behind why wave-based planning needs a live forecast, see the related article linked below.
Rolling wave planning in brief
Rolling wave planning is an established project-management technique: plan the work immediately ahead in detail, keep later work at a coarser level, and elaborate each subsequent "wave" as it approaches. The PMBOK describes the underlying principle as progressive elaboration. SAFe teams practice a form of it every Program Increment, planning the near PI to story level while later PIs remain at feature or epic level. The technique reflects the cone of uncertainty honestly: distant work cannot be specified accurately in advance.
What each wave does to a release forecast
Each elaboration wave changes the composition of the plan. Epics are split into stories, placeholders become concrete issues, and newly understood work is added to the Fix Version. When that happens, the app re-runs its Monte Carlo simulation over the team's actual throughput, and the committed date's confidence reading updates to reflect the release's current contents. The committed date is always carried with its confidence level (for example, "Nov 20 · 85%"), so the plan's growing accuracy and its effect on the date are visible together.
Using wave boundaries as checkpoints
Before elaborating a wave, note the committed date and its current confidence level.
Elaborate the wave in Jira as usual: split epics, add stories, assign the new issues to the release.
Read the recalculated confidence band. Movement caused by the new detail is reported in days.
If the P85 date now falls after the committed date, renegotiate while runway remains: reduce scope, adjust capacity, resequence a dependency, or re-set the commitment consciously.
Record the decision, and repeat at the next wave boundary.
Used this way, elaboration stops silently changing the meaning of the committed date; each wave ends with the team re-reading the commitment against current evidence.
Where it runs
Forecast and confidence calculations run natively inside Jira Cloud on Atlassian Forge. No data is exported to an external service to perform the calculation.
Related
Article: Rolling Wave Planning: The Waves Re-Plan the Work. The Date Needs to Keep Up.
Page: Live Confidence Bands: How They Keep a PI Confidence Vote Current
Page: Keeping a committed release date honest: scope, dependencies, and capacity
Article: The three things that erode a committed release date
Marketplace: Release Management, Roadmaps & Product Portfolio for Jira on the Atlassian Marketplace