Planning Poker cards explained
There's no one 'right' deck. The numbers you put on the cards shape how your team thinks about effort — so picking the deck is part of the work.
These are the decks you can pick when you start a session.
Scrum
The trimmed Fibonacci most Scrum teams reach for by default.
A Fibonacci deck without the very large numbers. The 0 and ½ stay so you can flag trivial work; the top end caps at 20, which is usually enough — anything bigger should probably be split before it goes into a sprint.
Fibonacci
The default for most agile teams.
Fibonacci puts gaps between the options on purpose. The bigger the work, the less anyone really knows — so the choices get further apart. You end up deciding "8 or 13?" instead of arguing about whether something's a 9 or a 10. The friction is the feature.
Modified Fibonacci
When you need to flag both trivial and oversized work.
Adds ½ for trivial-but-trackable work, and 40 and 100 at the top end. The 100 isn't an estimate — it's a stop sign that says "this is too big to estimate; break it up before we touch it." Respect the flag.
Sequential
Simple priority or risk scoring — not effort.
Linear 1–9 doesn't work well for effort because it invites endless single-point arguments. It does work for priority, risk, or confidence — anywhere you actually want a fine-grained scale and the values mean something ordinal rather than relative.
Half Card
Fine-grained estimation in half-points.
Half-point increments from 1 to 5½. Useful if your team genuinely splits hairs on small stories and you've decided that's a feature, not a bug. Most teams don't need this — Fibonacci's bigger gaps usually serve them better.
Power Of Two
For engineering teams that think in doublings.
Each card is double the last. Fits the way engineers reason about complexity — log scales, big-O, doubling cost as scope expands. The jumps get aggressive at the top, which is the point.
T-Shirt Size
Early refinement and leadership-level sizing.
Deliberately fuzzy — sometimes that's exactly what you want. T-shirts shine in early refinement ("do we even want this?") and at the portfolio level, where the conversation is about scope, not sprint capacity. The catch: T-shirts don't add up. "Twelve mediums in the sprint" doesn't mean anything about velocity.
The question mark
"I genuinely cannot estimate this." It's a stop sign. Usually the team's missing context or the story isn't defined enough yet — and that's a useful thing to surface before voting again.
The coffee cup
"I need a break." Take it seriously. Estimation fatigue tanks the quality of every round that follows, and a five-minute pause is cheaper than a bad sprint.
What ½ and 0 are for
Some decks include ½ and 0. Use ½ for trivial-but-trackable work ("rename this constant", "update the README"). Use 0 for stories that are already done — a useful reality check during refinement.
The 100 card
100 isn't an estimate. It's a flag. When someone plays a 100, they're saying "this is too big to estimate; break it up before we touch it." Respect the flag. Don't talk yourselves down to a 40 just to avoid the refinement work.
Custom decks
Plenty of teams roll their own. Risk scores for security review. Confidence levels (Low / Medium / High) for forecasting. Planning Poker is a pattern, not a fixed sequence — if your work needs a different scale, build one. (Need the why behind relative effort? Start here.)
Picking a deck for your team
Starting fresh? Use Fibonacci. It's the default for a reason: enough granularity for sprint planning, enough friction to keep arguments short. Switch to T-shirts only if you catch your team debating single-point differences. Move to Modified Fibonacci once you want to flag oversized stories explicitly.
Start a session and try a couple of decks back-to-back. Most teams know which one they want within two or three sprints.