Fibonacci, T-Shirt Sizes, or Powers of 2? Choosing a Planning Poker Scale
- Luke Garrigan
- Agile , Estimation , Scrum
- March 12, 2026
Table of Contents
When a team starts using planning poker, the first argument usually isn’t about a story—it’s about the cards. Should the deck be Fibonacci numbers? T-shirt sizes? Powers of 2? Here’s an honest comparison so you can pick one and move on.
Why Estimation Scales Aren’t Linear
The single most important idea: good estimation scales have growing gaps. You’ll almost never see a plain 1–10 deck, and there’s a good reason.
When a task is small, you can estimate it precisely—a 1 is clearly different from a 2. But as tasks get bigger, your uncertainty grows faster than the task does. Arguing whether something is a 20 or a 21 is a waste of everyone’s time. A scale with widening gaps forces the team to round to a meaningful bucket instead of pretending to a precision that doesn’t exist.
The Fibonacci Sequence (1, 2, 3, 5, 8, 13, 21…)
The default for most agile teams, and for good reason.
- ✅ Reflects real uncertainty — the gaps grow at roughly the right rate
- ✅ Forces a decision — there’s no “6”, so a story that feels between a 5 and an 8 makes the team talk it out
- ⚠️ Can feel abstract — new teams sometimes struggle with “what does an 8 even mean?”
Best for: most scrum teams doing story-point estimation on a backlog.
T-Shirt Sizes (S, M, L, XL, XXL)
Estimation stripped down to its simplest form.
- ✅ Zero math anxiety — non-technical stakeholders get it instantly
- ✅ Fast — great for high-level or early roadmap sizing
- ⚠️ Hard to track velocity — you can’t sum “three Mediums and a Large” without mapping sizes back to numbers
Best for: early backlog grooming, roadmap planning, or teams who find numbers cause more debate than they resolve.
Powers of 2 (1, 2, 4, 8, 16, 32)
A clean middle ground, popular with engineering-heavy teams.
- ✅ Intuitive doubling — “twice as big” is easy to reason about
- ✅ Wide gaps — like Fibonacci, it discourages false precision
- ⚠️ Coarser at the top — the jump from 16 to 32 is enormous
Best for: technical teams who like the mental model of work doubling in size.
So Which One Should You Use?
| Scale | Speed | Precision | Best for |
|---|---|---|---|
| Fibonacci | Medium | Good | Sprint story points |
| T-shirt | Fast | Low | Roadmap & early grooming |
| Powers of 2 | Medium | Good | Engineering-led teams |
The honest answer: it matters less than consistency. Whichever scale you pick, the value comes from the team using the same scale long enough to build a shared sense of what a “5” feels like. Switch decks every sprint and your velocity numbers become meaningless.
If you’re unsure, start with Fibonacci—it’s the industry default, so new hires already know it.
Try All Three for Free
Planfree.dev ships with Fibonacci, T-shirt, and powers-of-2 decks built in—switch between them in the room settings with one click, no account required. Spin up a room, try a scale on a few real stories, and see what your team gravitates toward.
Planfree.dev is free, open-source, and has no participant limits. Contribute or suggest a deck on GitHub.