Rules exist for reasons. But blind adherence to rules without understanding their purpose makes you a rule-follower, not a practitioner. Here are planning poker's official rules, why they exist, and when it's actually smart to bend them.
Some rules are sacred (simultaneous reveal). Others are negotiable based on your team's context (Fibonacci vs T-shirt sizing). Knowing the difference is what separates cargo-cult agile from thoughtful practice.
The Official Rules
Use Fibonacci Sequence
Standard cards are: 0, 1, 2, 3, 5, 8, 13, 21, ?, ☕. The growing gaps reflect increasing uncertainty in larger estimates.
Why: Wider gaps prevent false precision. You can't meaningfully differentiate between 17 and 18 story points.
Everyone Votes Simultaneously
All participants select and reveal their cards at the same time. No one shows their estimate before others.
Why: Prevents anchoring bias. If the tech lead goes first, everyone unconsciously adjusts toward their number.
Only Implementers Vote
Only team members who will actually do the work cast estimates. Product owners, stakeholders, and managers observe but don't vote.
Why: Those doing the work have the best context. Non-implementers voting dilutes accuracy and creates resentment.
Discuss Outliers Before Re-Voting
When estimates vary widely, highest and lowest voters explain their reasoning. Then the team votes again.
Why: The discussion is where learning happens. One person knows about technical debt, another about a reusable component.
No Averaging
Don't calculate the mathematical average. If votes are 3, 5, and 8, you pick 5 or 8—not 5.33.
Why: Fibonacci is intentionally non-linear. Averaging destroys the scale and produces meaningless precision.
Consensus Over Speed
Keep discussing and re-voting until estimates converge or the team agrees the story needs breaking down.
Why: Rushing to a number without alignment means the team doesn't share understanding of the work.
Estimate Relative Complexity, Not Time
Story points measure effort/complexity, not hours. A 5-point story is roughly half the effort of an 8-point story.
Why: Time estimates encourage micro-management and don't account for interruptions, meetings, or unknowns.
Question Mark Means Not Enough Info
The ? card signals "I can't estimate this." It's not a cop-out—it means the story needs clarification.
Why: Forcing estimates on vague stories leads to wildly inaccurate planning. The ? card protects against this.
Common Variations
These aren't rule breaks—they're adaptations teams use successfully. Choose based on your team's context, not because they're trendy.
T-Shirt Sizing
What Changes
Use XS, S, M, L, XL instead of Fibonacci numbers
When to Use
Early roadmap planning or stakeholder communication. Less intimidating than numbers.
Tradeoffs
Less precision than Fibonacci. Harder to track velocity mathematically.
Modified Fibonacci
What Changes
Add 1/2 card for tiny tasks, or use 40/100 for enormous epics
When to Use
Teams with lots of micro-tasks or dealing with multi-sprint initiatives.
Tradeoffs
More cards = more choice paralysis. Can encourage over-granularity.
Powers of 2
What Changes
Use 1, 2, 4, 8, 16, 32 instead of Fibonacci
When to Use
Engineering teams who think in binary, or when you want simpler math.
Tradeoffs
Gaps grow faster than Fibonacci. Less middle ground between estimates.
Async Voting
What Changes
Team members submit estimates in a tool over hours/days instead of live session
When to Use
Fully distributed teams across time zones, or when live meetings are impossible.
Tradeoffs
Loses the real-time discussion. Slower iteration on unclear stories.
Silent First Round
What Changes
First vote happens without any discussion. Second round allows questions.
When to Use
When you want to see raw gut reactions before context influences estimates.
Tradeoffs
Can waste time if the story was genuinely unclear from the start.
When to Break Rules
Rule
Only implementers vote
When to Break It
QA/testers should vote if testing complexity is significant
Example
Story involves complex cross-browser testing or security validation. Testers know the effort better than devs.
Rule
Use Fibonacci sequence
When to Break It
Brand new team unfamiliar with story points
Example
Try T-shirt sizing for the first 2-3 sprints, then migrate to Fibonacci once team understands relative estimation.
Rule
No averaging
When to Break It
Never. Seriously, don't do this.
Example
Even when estimates are 5 and 8, pick one. If you must compromise, choose the higher estimate for safety.
Rule
Everyone votes simultaneously
When to Break It
Literally never. This destroys the entire technique.
Example
Sequential voting reintroduces all the biases planning poker was designed to eliminate.
Rule
Consensus over speed
When to Break It
When 80% of team agrees and dissenters can't articulate specific concerns
Example
Seven people vote 5, one person votes 8 but says "just feels bigger." Go with 5 or split the difference to 8 for safety.
Rule
Estimate complexity, not time
When to Break It
When leadership demands time-based commitments
Example
Track historical data: if your 5-point stories average 1.5 days, you can translate. But internally, keep estimating in points.
House Rules Worth Trying
These aren't official, but teams report they improve estimation quality or team dynamics. Experiment and keep what works.
The "Lunch Card" Override
If someone plays the coffee card and 60%+ of the team agrees, estimation stops immediately for a break.
Why It Works
Prevents zombie estimation. A tired team produces garbage estimates. Better to pause than push through.
When to Adopt
Long estimation sessions (2+ hours) or when team energy is clearly fading.
Three Strikes Split
If a story requires more than three rounds of voting without consensus, it gets auto-flagged for breakdown.
Why It Works
Prevents estimation paralysis. If you can't agree after three tries, the story is too vague or too big.
When to Adopt
Teams that debate endlessly or struggle to break down large stories.
Newbie Votes First
Newest team member must commit to an estimate before discussion starts. Forces them to engage.
Why It Works
Juniors often stay silent to avoid looking wrong. This rule makes their perspective heard.
When to Adopt
Teams with new members who are hesitant to speak up.
No Phone Poker
Anyone on their phone during story explanation must play the ? card—they can't estimate what they didn't hear.
Why It Works
Enforces attention. You can't estimate a story you half-listened to while checking Slack.
When to Adopt
Remote teams with attention drift, or in-person teams with distraction issues.
Wildcard Wednesday
Once per sprint, team can vote to use a completely custom card value if standard Fibonacci doesn't fit.
Why It Works
Acknowledges that sometimes work doesn't fit the mold. Pressure valve for edge cases.
When to Adopt
Teams dealing with unusual work types that resist standard estimation.
Document Outlier Wisdom
When someone's outlier estimate catches a major risk, note it in sprint retro. Track who spots hidden complexity.
Why It Works
Celebrates good estimation. Reinforces that variance is valuable, not a problem to eliminate.
When to Adopt
Teams that rush to consensus too quickly or dismiss outliers without discussion.
The Bottom Line
Rules should serve your team, not the other way around. Understand why each rule exists, then decide which ones apply to your context. The simultaneous reveal is sacred—it's the core mechanism. Everything else is negotiable.
The best teams know the rules well enough to break them intelligently. The worst teams either follow rules blindly or ignore them completely. Be thoughtful about which camp you're in.
Play by Your Rules
Create a room and establish your team's house rules. Start with the basics, evolve as you learn.
Create Free Room