Fibonacci in planning poker
The Fibonacci sequence is the backbone of planning poker. Understanding why it works helps teams use it better, avoid common pitfalls, and choose the right scale for their context.
What is the Fibonacci sequence?
The Fibonacci sequence is 1, 1, 2, 3, 5, 8, 13, 21, 34… where each number is the sum of the two preceding ones. In planning poker, the sequence is typically shortened to 1, 2, 3, 5, 8, 13, 21 plus a ? card.
The key property is that gaps grow with the size of the number. Between 1 and 2, the gap is 1. Between 13 and 21, it is 8. This is not arbitrary, it reflects how human cognition works when estimating uncertainty.
Why Fibonacci instead of 1–10?
A linear scale (1, 2, 3, 4, 5, 6, 7, 8, 9, 10) creates a false sense of precision. Is a task a 6 or a 7? Teams spend time debating distinctions that carry no real information.
With Fibonacci, the choice is always coarser: is this a 5 or an 8? That is a meaningful difference. The forced coarseness prevents analysis paralysis and pushes teams towards discussion rather than point arithmetic.
Psychologists call this subitizing: humans can reliably distinguish small quantities at a glance, but lose accuracy quickly beyond 5–7 items. Fibonacci aligns the scale with this cognitive limit.
The ? card: an underused signal
The ? card is one of the most valuable in the deck. It means: "I cannot estimate this story." Valid reasons include:
- The acceptance criteria are unclear.
- The story depends on an unresolved technical decision.
- The story is too large to estimate as a whole (needs splitting).
- The team member lacks the context to judge complexity.
A ? vote should always pause the estimation and trigger a discussion. It is a quality gate, not a failure.
Standard vs. modified Fibonacci
| Scale | Values | Best for |
|---|---|---|
| Standard | 1, 2, 3, 5, 8, 13, 21, ? | Most Scrum teams, sprint-level estimation |
| Modified | 0, 0.5, 1, 2, 3, 5, 8, 13, 20, 40, 100, ? | Teams with very small tasks or large epics in the same backlog |
| Powers of 2 | 1, 2, 4, 8, 16, 32, 64, ? | Technical teams where complexity doubles predictably |
How to pick your card
Good estimators think in reference tasks. Before your first session, identify 2 or 3 anchor stories the team knows well: one that is clearly a 1, one a 5, one a 13. Every new story is then compared to these anchors rather than estimated in isolation.
Ask yourself:
- Is this story simpler or more complex than our 5-point anchor?
- Does it involve more unknowns than our 8-point anchor?
- Could one developer finish it in a single day? (Usually a 1 or 2.)
Frequently asked questions
Why does planning poker use Fibonacci numbers?
Fibonacci numbers model growing uncertainty: the larger a task, the less precisely it can be estimated. The increasing gap between values forces the team to take a clear position without pretending to false precision.
What does the ? card mean in planning poker?
The ? card signals "I cannot estimate this task." It is a healthy indicator that the story lacks clarity, is too large, or has unresolved dependencies. It triggers a clarification discussion before any estimate is locked in.
Should I use standard or modified Fibonacci?
Start with standard (1, 2, 3, 5, 8, 13, 21). Add 0 and 0.5 only if your team regularly debates whether a task is "almost nothing." Add values above 21 only if you estimate epics alongside sprint stories. Adding complexity rarely helps, simplicity keeps velocity reliable.
Is there a correct Fibonacci value for a given task?
No. Story points are relative to your team's reference frame, not to clock time. What matters is consistency: a 5 today should feel as complex as a 5 six months from now. That consistency is what makes velocity a useful forecasting tool.
When should a team switch to T-shirt sizes?
T-shirt sizes are better for rough backlog grooming at the start of a project or for non-technical stakeholders. Once the team needs velocity tracking, switch to Fibonacci story points for numerical tracking over sprints.