Key Takeaways
- Hofstadter’s Law says work always takes longer than you expect, even when you allow for Hofstadter’s Law. It names a tendency. It does not say how much longer, or why.
- A team’s confidence vote is expert engineering judgment on the plan in front of it. The lateness that follows usually comes from changes made after the vote, not from a bad vote.
- Three things move a committed release date after the vote: scope creep, dependencies surfaced too late, and capacity changes.
- Replace “it always takes longer” with a number: the gap in days between the committed date and its live 50/85/95% confidence band, with the cause attached.
The release review is going badly. The date the team committed to in planning was 30 May. It is now 14 May, and the board says early June at best. Someone in the room says, half-joking, “Hofstadter’s Law.” A few people laugh. The meeting moves on to who should email the customer.
The joke lands because everyone has lived it. But it also closes the conversation at the exact point where the useful question starts: longer by how much, and because of what?
What Hofstadter’s Law actually says
Douglas Hofstadter wrote it in his 1979 book Gödel, Escher, Bach: “It always takes longer than you expect, even when you take into account Hofstadter’s Law.” The law is self-referential on purpose. Add a buffer because you know things run late, and the work still runs past the buffer. Add a buffer for the buffer, and the same thing happens.
Software teams quote it because it matches experience. It sits alongside the planning fallacy, the research finding that people underestimate their own task durations even when they know similar tasks ran late. Both describe a real pattern. Neither tells a delivery lead what to do on 14 May.
The vote was not the problem
It is tempting to read Hofstadter’s Law as a verdict on the team: the estimate was wrong, so the people who made it were optimistic. That reading is usually unfair. When a team votes four fingers on a plan at the end of planning, it is giving expert engineering judgment on the scope, dependencies and capacity that existed in that room on that day. The engineers who will build the work know more about it than anyone else.
The trouble is that the plan they voted on stops existing within days. A confidence vote is a snapshot. The release keeps moving after the snapshot is taken, and the vote has no way of knowing. And once the date itself becomes the target, status reports start protecting it instead of measuring it, which is Goodhart’s Law at work.
Where the “longer” comes from after commitment
Look at any release that slipped and trace the changes after the commitment. The same three causes appear again and again. They are the three things that erode a committed release date:
- Scope creep. A stakeholder asks for one more field. A story splits in two during refinement. A bug found in testing goes into the same Fix Version. Each addition is small; together they change the remaining work.
- Dependencies surfaced too late. A platform team changes an API. A vendor delivery slips. A blocking link appears in week four that nobody could have drawn in planning.
- Capacity changes. An engineer is pulled onto an incident. A new joiner is still ramping. A public holiday or two weeks of PTO was never entered.
This is the recursion in Hofstadter’s Law, made concrete. A buffer added at planning protects against the uncertainty the team could see. It cannot protect against changes that had not happened yet. Every one of those changes is recorded in Jira the day it happens. What is missing is its cost to the committed date.
From “always longer” to “eight days longer, because of scope”
The way out of the joke is to put a number on it, continuously. Instead of a single date, the team commits to a date with a confidence band: the 50%, 85% and 95% dates from a Monte Carlo forecast over its own throughput. Then the band keeps updating from Jira after the commitment.
An illustration: a release is committed for 30 May at 85% confidence. In week two, eleven stories are added to the Fix Version; the 85% date moves to 3 June. In week three, a blocking dependency on another team appears; the 85% date moves to 7 June. The gap is now eight days, split between scope and a dependency, and the team knows it with three weeks to go. That is enough time to move two “should have” stories out, re-sequence around the dependency, or reset expectations with the customer while it is still a conversation rather than an apology.
Over several releases, the same record answers a second question: does the team’s 85% date land about 85% of the time? That is how forecast accuracy should be judged, by calibration across releases rather than by one missed date.
How Advanced Release Planning keeps the date honest
Advanced Release Planning for Jira does not estimate issues or replace the team’s vote. The team still owns the assumptions: which throughput window to trust, whose capacity is in, what is in scope. The app does the arithmetic at Monte Carlo scale and keeps it current.
- The committed date carries a live confidence band at 50%, 85% and 95%.
- Work added to the Fix Version after the commitment baseline counts as new scope.
- Dependency-aware forecasting reads blocking Jira links, so a late dependency moves the forecast.
- Capacity changes, non-working days and PTO flow into available throughput.
- Drift is reported in days and attributed to scope, dependencies or capacity.
Setup steps are in the Hofstadter’s Law & the Committed Release Date reference.
Hofstadter’s Law will keep getting a laugh in release reviews. It does not have to be the end of the discussion. A confidence vote is a snapshot; Advanced Release Planning keeps it live.
FAQ
What is Hofstadter’s Law?
An adage from Douglas Hofstadter’s 1979 book Gödel, Escher, Bach: “It always takes longer than you expect, even when you take into account Hofstadter’s Law.” It is widely quoted in software and project management to describe how schedules overrun even after buffers are added.
Does adding more buffer beat Hofstadter’s Law?
Not on its own. A buffer set at planning covers the uncertainty visible then. Scope, dependency and capacity changes after the commitment need to be measured as they happen, in days against the committed date.
See your committed date and its live confidence band in Advanced Release Planning, or try it free on the Atlassian Marketplace.




1 Comment
Leave your reply.