Two popular approaches, two different philosophies. Which one fits your team? Both are covered, alongside other methods, in our guide to agile estimation techniques for sprint planning.
What is the difference between Fibonacci and T-shirt sizing?
Fibonacci and T-shirt sizing are two relative estimation scales used by agile teams. Fibonacci estimation assigns story points taken from a modified Fibonacci sequence (1, 2, 3, 5, 8, 13, 21). Because these numbers can be added up, the team can measure its velocity and forecast how many sprints a backlog will take. T-shirt sizing sorts work into clothing sizes (XS, S, M, L, XL, XXL). It is faster to learn, easier for non-technical stakeholders to grasp and well suited to rough roadmap estimates, but sizes cannot be summed, so they cannot feed a velocity. In short, Fibonacci fits sprint planning for stories the team is about to build, while T-shirt sizing fits early, high-level estimation of epics and features. Many teams combine both: T-shirt sizes for the long-term backlog, then Planning Poker with Fibonacci cards once stories approach a sprint.
How do Fibonacci and T-shirt sizing compare?
| Criterion | Fibonacci | T-Shirt Sizing |
|---|---|---|
| Granularity | Fine: 7 or 8 values with growing gaps | Coarse: 5 or 6 sizes |
| Best use | Sprint planning, refined user stories | Roadmap, discovery, epics and features |
| Velocity tracking | Yes, points add up across a sprint | No, sizes are not additive |
| Long-term backlog | Slow on dozens of vague items | Fast way to size a large backlog |
| Anchoring risk | Numbers invite "8 points = 8 days" thinking | Each person reads sizes differently |
| Learning curve | A few sprints to calibrate | Immediate |
| Stakeholder friendliness | Needs explaining | Intuitive for everyone |
| Debates generated | Sometimes intense ("5 or 8?") | Usually quick |
Neither scale removes anchoring on its own. The real protection is a simultaneous vote, where everyone picks a value privately before anyone speaks. That is the core rule of Planning Poker, and it works with both scales.
When should you use Fibonacci?
Fibonacci allows more nuanced distinctions. The gap between a 5 and an 8 is meaningful, and that granularity helps a team decide exactly what fits into a two-week sprint.
Use Fibonacci when:
- stories are refined, with acceptance criteria the team understands;
- you want to track velocity over time and forecast releases;
- the team is stable, so its scale keeps the same meaning from sprint to sprint.
The growing gaps are not arbitrary. Mike Cohn links them to Weber's law: we perceive differences as percentages, and Fibonacci values keep a roughly constant relative gap. Read more about why agile teams use Fibonacci.
When should you use T-shirt sizing?
T-Shirt Sizing shines in exploratory phases. When you're discovering a new product, building a quarterly roadmap or bringing non-technical stakeholders into the conversation, clothing sizes are more intuitive. You don't need to explain Fibonacci to a marketing director: everyone knows what an XL is.
Use T-shirt sizing when:
- the backlog holds dozens of items that are not yet split into stories;
- you need a first ranking of features by size, in an hour rather than a day;
- the team is new to relative estimation and needs a gentle start.
Mike Cohn sees ease of getting started as the main advantage of T-shirt sizes. He also points out their two limits: sizes are not additive, and "your view of an XL may not match mine". Our T-Shirt Sizing guide explains how to set it up.
Can you convert T-shirt sizes to story points?
Yes, as long as each team defines its own mapping and treats it as a rough bridge, not an exact conversion. Cohn recommends putting underlying numbers on sizes (for example, Medium = 5) and then gradually switching to the numbers themselves. Here is a typical starting point:
| T-shirt size | Example story points | What it usually means |
|---|---|---|
| XS | 1 | Small change, well understood |
| S | 2 or 3 | Simple story, few unknowns |
| M | 5 | Standard story, some uncertainty |
| L | 8 | Large story, worth a risk check |
| XL | 13 | Very large, consider splitting |
| XXL | 21 or more | Epic, must be split before a sprint |
The mapping is a convention, not a law: Cohn's own example sets a medium at 5 and a large at 10. What matters is that the team writes its mapping down and applies it consistently. Without numbers, sizes cannot be added: you cannot tell stakeholders that a release is "three mediums and two larges" away.
Can you combine the two methods?
Yes, and it's even a good practice. The two scales work at different horizons:
- Roadmap and discovery: T-shirt sizes for epics and features, to compare options and order the backlog.
- Backlog refinement: as an item gets closer, the team splits it into stories and gives each one a Fibonacci estimate.
- Sprint planning: only stories with a Fibonacci estimate of 13 or less enter the sprint, and velocity tells you how many fit.
This is close to James Grenning's original intent: his 2002 Planning Poker paper was written for release planning, where precision of individual estimates matters less than seeing the overall scope.
Which method should your team pick?
- You track velocity and plan sprints: Fibonacci.
- You build a roadmap or size a fresh backlog: T-shirt sizing.
- Your team is new to estimation: start with T-shirt sizes, map them to numbers, then switch to Fibonacci.
- Stakeholders join the session: T-shirt sizes keep the conversation accessible.
Rule of thumb: pick the method your team understands and actually uses. The best estimation technique is the one that's applied with discipline, not the one that's theoretically perfect.
Frequently asked questions
Is T-shirt sizing less accurate than Fibonacci? T-shirt sizing is deliberately less precise than Fibonacci, which is a feature rather than a flaw. With only five or six sizes, it gives a quick, coarse picture of a large backlog. Fibonacci offers finer distinctions that suit refined stories, but that extra precision only helps once the team understands the work well enough.
Can you calculate velocity with T-shirt sizes? Not directly. T-shirt sizes are labels, so they cannot be added up across a sprint, and a velocity is a sum. To track velocity with T-shirt sizing, a team must first map each size to a number, such as Medium = 5, and then sum those numbers. At that point, it is effectively using story points.
Which estimation method is better for beginners? For teams new to agile estimation, T-shirt sizing is usually the easier entry point: everyone understands S, M and L without training. Once the team is comfortable comparing items relative to each other, mapping sizes to numbers and moving to the Fibonacci scale unlocks velocity tracking and more reliable sprint forecasts.
Can you play Planning Poker with T-shirt sizes? Yes. Planning Poker works with any scale, including T-shirt sizes. Each voter picks a size privately, everyone reveals at once, and the team discusses the gaps. Planning Poker Online offers Fibonacci, T-shirt (XS to XXL) and powers-of-2 scales, chosen by the Scrum Master for each session.
Sources
- Mike Cohn, Estimating with Tee Shirt Sizes, Mountain Goat Software
- Mike Cohn, Why the Fibonacci Sequence Works Well for Estimating, Mountain Goat Software
- Mike Cohn, What Are Story Points?, Mountain Goat Software
- James Grenning, Planning Poker or How to avoid analysis paralysis while release planning, 2002
Estimate with the scale that fits
Whichever scale you choose, create a free Planning Poker session: Fibonacci and T-shirt sizes are both available, and your team can vote in seconds. See all the features.
