Every release commitment has a number hiding behind it. Not the date — the confidence in the date. And most teams never say it out loud.

Picture the end of PI planning. The room settles on a release date, someone writes it on the board, and the fist-of-five confidence vote comes back solid. Everyone commits. But commit to what, exactly? The date the team could hit if everything broke right? The date they’d hit four times out of five? Nobody asked. The date got committed; the confidence behind it stayed invisible.

That invisible number is the whole difference between a commitment you can defend and one that quietly comes apart. So let’s make it visible — and make choosing it a deliberate act.

The confidence you commit at is a decision, not a default

When Advanced Release Planning for Jira forecasts a release date, it doesn’t hand back one date. It hands back a band — three dates at three confidence levels, run by Monte Carlo simulation against your team’s real throughput (the same past sprint velocity you can use to forecast a whole quarter):

• P50 — you’d hit this date about half the time. A coin flip. Fine as an internal target, dangerous as an external promise.
• P85 — you’d hit it roughly five times out of six. This is the level most SAFe teams treat as committable, and the app’s default on-track threshold. It’s the data equivalent of a confident fist-of-five.
• P95 — a conservative, high-certainty read. The right call when a miss is expensive: a regulated launch, a contract deadline, or a launch you’ve already advertised.

Choosing between them is expert judgment, not arithmetic. The simulation places each date on the calendar, but which one you commit to depends on things only the team knows: how costly a miss is, who’s counting on the date, how much of the scope is still soft. The tool does the math at a scale no spreadsheet can; the team owns the call. That’s the honest version of the SAFe confidence vote — run on evidence the team validates, not a show of hands taken on faith.

If that split sounds familiar, it’s because SAFe already works this way. Committed objectives are the ones a team is confident enough to promise; stretch objectives sit below the line as deliberate buffer. A confidence band is the same idea drawn on a calendar — P85 is where you commit, and the space between P50 and P85 is your stretch: the margin you’re choosing to hold in reserve.

How to actually pick a level

A simple heuristic: match the confidence level to the cost of being wrong.

• If a miss slips an internal demo by a week and annoys no one, commit at P50 and move fast.
• If a miss means a hard conversation with a customer or an executive, commit at P85. It’s the honest default for most release commitments — enough margin to absorb normal variance without padding the plan into fiction.
• If a miss triggers penalties, compliance exposure, or a launch you can’t move, commit at P95 and defend it.

The mistake isn’t picking the “wrong” level. The mistake is not picking at all — inheriting whatever date a roadmap tool typed in, with no idea whether it’s a coin flip or a near-certainty. Once you’ve chosen, everyone reading the plan sees not just the date but the odds behind it. (If you’re the one being handed the number, here’s how to read that percentage in a review.)

A confidence level is a snapshot — keep it live

Here’s the part that catches teams out. You choose P85, you commit, you feel good. Then the eight to twelve weeks of the PI actually happen.

Scope creeps in. A dependency nobody flagged surfaces in week five. Two engineers get pulled onto an incident. None of it shows up in a vote you took once, on day one — but every bit of it eats the margin that made 85% mean something. The date on the board hasn’t moved. Your real confidence in it has. These are the three things that quietly erode a committed release date, and they don’t wait for your next planning session.

This is why a one-time confidence decision isn’t enough. Advanced Release Planning recalculates the confidence continuously, tracking those same three inputs — scope, dependencies, capacity — against the committed date, in days, in real time. When your 85% quietly slides to 62%, you see it while there’s still room to renegotiate scope or move the date on purpose — not in the retro, explaining a miss. See how the confidence band works in the docs.

A confidence level you choose once is a promise you hope holds. A confidence level kept live is a promise you can manage. Same number — but one tells you the truth all the way to release.

Commit on data, not hope

See the confidence band on your own release dates, choose the level you commit to on real throughput, and watch it stay honest as the work shifts. Try Advanced Release Planning for Jira, free for 30 days.

Rob Borges

Rob Borges spent the first half of his career hunting waste as a Lean Six Sigma Master Black Belt and wrangling portfolios as a PMI Portfolio Manager — until he realized the tidiest process on earth still can't out-plan an optimistic human. So he switched sides and became an Agile Coach. These days he helps Kanban, Scrum, and Agile-at-Scale teams trade gut-feel deadlines for flow data, turning "we'll probably ship sometime in Q3" into release dates leaders can actually take to the board. His rule of thumb: hope is not a forecast, but your throughput history usually is.

See Full Bio