Most of a planning poker deck is numbers, and most teams only ever play the numbers. The four cards that are not numbers are the ones that keep a session honest.
The standard deck
- 0
- ½
- 1
- 2
- 3
- 5
- 8
- 13
- 21
- 34
- ?
- ∞
- ☕
Every participant holds an identical deck. On each story, everyone picks one card face down, and the whole set is turned over at the same moment. Decks vary — some stop at 21, some round the top end to 20, 40 and 100 — but the structure is consistent: fine steps at the bottom, widening steps at the top, plus a few cards that say something other than a size.
The numbered cards
| Card | Meaning |
|---|---|
| 0 | Already done, or too small to be worth tracking. Occasionally used for work that ships as a side effect of another story. |
| ½ | Real but trivial. Not in every deck — teams that track very small work tend to want it, teams that batch small work do not. |
| 1–3 | Small and well understood. The routine end of the backlog. |
| 5–8 | Substantial work. An 8 is worth a second look to see whether it splits cleanly. |
| 13–34 | Large and usually under-understood. Treat these as a refinement signal rather than a plan. |
The cards that are not sizes
? — "I cannot estimate this"
The question mark means the player does not understand the story well enough to put a number on it. This is genuinely valuable information and it is routinely suppressed, because playing it feels like admitting ignorance in front of the team.
A facilitator's job is to make it safe. One question mark usually means a missing piece of context that a sentence from the product owner will fix. Several means the story is not ready, and no amount of voting will make it ready. Either way the alternative — a guessed number that gets averaged in as though it carried information — is strictly worse.
∞ — "Too big to estimate"
Infinity is not a very large number; it is a refusal. The player is saying the item is an epic, not a story, and cannot meaningfully be sized as one unit. The response is to split it, not to settle on 34 as a compromise.
☕ — "I need a break"
The coffee cup is the card people assume is a joke. It is not. Estimation quality drops noticeably after about an hour of concentration, and the last thirty minutes of a long session produce votes that are mostly fatigue. A facilitator who honours the coffee cup gets better numbers for the rest of the session.
GiveItPoints keeps this as a first-class vote rather than a gimmick — an abstain option that records the player as present and deliberately not sizing, so the statistics are not skewed by treating it as a zero.
If nobody on your team has ever played a question mark, that is not because every story has been clear. It is because playing one does not feel safe yet.
Physical cards or a shared tool
Physical decks work well in one room, and there is something to the ritual of turning cards over together. They fail the moment anyone dials in. Holding a card to a webcam is neither simultaneous nor private: the person with the best connection reveals first, everyone else sees it, and the anchoring the method exists to prevent walks straight back in.
A shared tool enforces the reveal mechanically, which matters more than it sounds. It also gives you the thing physical cards never could — a record. Knowing that the team sized this kind of story at 5 four times and it took longer every time is the sort of feedback that actually improves estimates.
Choosing your own values
Nothing obliges a team to use Fibonacci. Powers of two (1, 2, 4, 8, 16) give a similar widening effect. T-shirt sizes drop numbers altogether. Some teams cap the deck at 8 on the principle that anything larger should have been split before the meeting — a strict rule, and a surprisingly effective one.
GiveItPoints lets a facilitator define the deck per team, numeric or t-shirt, so the scale matches how the team actually works rather than the other way round. See custom voting scales for how that works in practice.
Common questions
How many cards should a deck have?
Eight to ten numeric values covers almost every team. More than that and the extra options slow the vote without improving it.
What happens if someone plays the coffee cup?
Take the break. The vote for that story is retaken afterwards. A facilitator who overrides it teaches the team the card is decorative.
Can the product owner hold a deck?
They can hold one, but they should not vote. The estimate belongs to the people doing the work; a product owner's number carries an implicit expectation that distorts the round.