Understanding Fibonacci Sequence Estimation
The Fibonacci sequence—1, 2, 3, 5, 8, 13, 21, 34—provides the foundation for most planning poker implementations. Each number equals the sum of the previous two numbers, creating progressively larger gaps that reflect increasing uncertainty as work grows more complex.
Teams using Fibonacci estimate stories by assigning one of these numbers as story points. A login form might be a 3, building an API integration an 8, and architecting a new microservice a 21. The numbers are deliberately non-linear because our ability to distinguish complexity diminishes as work grows larger.
The mathematical elegance appeals to analytically-minded teams. The sequence appears in nature, from nautilus shells to flower petals, suggesting some fundamental truth about growth and complexity. More practically, it prevents the false precision of linear scales.
Teams track velocity by summing completed story points per sprint. If you finish stories totaling 32 points across two weeks, your velocity is 32. This number becomes the baseline for future sprint planning, enabling predictable commitments.
Understanding T-Shirt Sizing
T-shirt sizing uses familiar categories: XS, S, M, L, XL, and sometimes XXL. Instead of abstract numbers, teams estimate with sizes everyone intuitively understands from buying clothing. A simple bug fix is XS, a standard feature is M, and a complex integration is XL.
The approachability matters. New team members grasp T-shirt sizing immediately, while Fibonacci requires explanation. Non-technical stakeholders find sizes less intimidating than story points, reducing the communication friction that often plagues agile teams.
Some teams use T-shirt sizes directly for planning. They might complete 2 XL, 3 M, and 5 S stories per sprint and forecast based on those categories. Others convert sizes to numeric values for velocity calculation: XS=1, S=2, M=3, L=5, XL=8, effectively creating Fibonacci in disguise.
The flexibility creates both opportunity and confusion. Without clear conversion rules, T-shirt sizing can become mushy. One person's large differs from another's, more dramatically than numeric point disagreements.
Precision and Granularity Comparison
Fibonacci provides more granularity in the mid-range where most stories fall. Between small and large work, you have clear options: 3, 5, 8. This supports nuanced estimation when a story feels bigger than small but not quite medium-large.
T-shirt sizing offers fewer gradations. Between S and L, you only have M. Some stories clearly aren't medium but determining whether they're small-ish medium or medium-ish large becomes difficult. Teams sometimes create intermediate sizes (S/M) which defeats the simplicity advantage.
The granularity affects velocity stability. Fibonacci's finer gradations produce more consistent sprint-to-sprint velocity because individual estimation differences average out. T-shirt sizing's coarser buckets create more variability—swapping one XL story for two S stories dramatically changes the total even if actual effort is similar.
However, Fibonacci's precision can be false. The difference between 5 and 8 points feels concrete, but it often represents the same uncertainty as the difference between M and L shirts. The numeric specificity creates an illusion of accuracy that doesn't reflect reality.
Learning Curve and Adoption Speed
T-shirt sizing wins overwhelmingly on initial adoption. Explain sizes once, show a few examples, and teams start estimating productively within hours. The metaphor requires no mathematical background and works across cultures and experience levels.
Fibonacci requires more upfront investment. Teams need to understand why the sequence matters, how it reflects uncertainty, and why we skip numbers like 4, 6, 7. New members ask: "Why can't we use 10 instead of 8 or 13?" The explanation is valid but adds cognitive overhead.
That said, Fibonacci teams often develop deeper estimation sophistication faster. The structured scale encourages rigorous thinking about complexity. T-shirt teams sometimes become sloppy, treating sizing as a quick gut-check rather than collaborative analysis.
For distributed teams using tools like FreeScrumPoker, both methods work equally well technically. The choice impacts how quickly remote team members align on what estimates mean, with T-shirts providing faster initial calibration.
Velocity Tracking and Forecasting
Fibonacci supports precise velocity tracking out of the box. Sum completed points, plot over time, calculate rolling averages. The numeric nature enables straightforward statistical analysis: standard deviations, trend lines, confidence intervals.
Teams reach stable, predictable velocity faster with Fibonacci. Research shows Fibonacci teams achieve consistent velocity within 3-4 sprints, while T-shirt teams converting to numbers take 5-6 sprints to stabilize. The structured scale accelerates calibration.
T-shirt sizing requires conversion decisions for quantitative forecasting. Converting to Fibonacci values works but raises the question: why not use Fibonacci directly? Keeping sizes as-is means tracking "we complete about 1 XL, 2 L, 3 M, and 5 S stories per sprint." This is harder to communicate and aggregate across epics.
For teams practicing #NoEstimates or pure count-based forecasting, T-shirt sizing's disadvantages in velocity tracking matter less. If you're counting stories regardless of size, the estimation scale becomes nearly irrelevant beyond ensuring stories are similarly sized.
Stakeholder Communication
Stakeholders generally find T-shirt sizes more intuitive than story points. "This epic has three large features and five medium features" communicates more clearly to business partners than "this epic is 65 story points."
However, sizes can backfire when stakeholders apply real-world logic that doesn't transfer. "If you can do a large in two days, why does this large take five days?" The variability within size categories confuses people expecting consistency, similar to how actual clothing sizes vary dramatically between brands.
Fibonacci's abstraction creates healthy distance. When you explain "story points measure relative complexity using a standardized scale specific to our team," stakeholders accept that 8 points means different things in different contexts. The non-intuitive nature paradoxically reduces misunderstanding.
Roadmap planning typically works better with T-shirt sizing. Executives can grasp "Q1 will focus on these five large initiatives" more easily than "Q1 will deliver 180 story points." The size metaphor maps to strategic thinking patterns better than abstract numbers.
Team Collaboration Dynamics
Planning poker with Fibonacci encourages healthy debate. When estimates diverge—one person says 3, another 8—the gap is obvious and demands discussion. This surfaces hidden assumptions and increases shared understanding.
T-shirt sizing can allow too easy consensus. "Is this medium or large?" "Eh, medium I guess." Without the structure of numeric gaps, teams sometimes settle on estimates without deep analysis. The casual nature that aids adoption can undermine estimation rigor.
Conversely, Fibonacci can create destructive precision arguments. Teams waste time debating 5 versus 8 points when the difference doesn't meaningfully affect planning. T-shirt sizing's coarseness prevents bikeshedding—nobody argues whether something is medium versus medium-large when only five sizes exist.
Cross-functional teams with wide skill variance often prefer T-shirts. Junior developers, designers, and product managers find sizes more accessible than points. This inclusivity improves estimation quality by capturing diverse perspectives rather than deferring to technical experts.
When to Choose Fibonacci
Choose Fibonacci when you need:
- Precise velocity tracking: Teams optimizing for predictable delivery timelines benefit from Fibonacci's granularity and statistical properties.
- Sprint planning focus: If estimation primarily serves sprint planning rather than roadmap planning, Fibonacci's numeric precision adds value.
- Experienced agile teams: Teams with 2+ sprints experience who've outgrown simpler methods gain from Fibonacci's structure.
- Mathematical inclination: Engineering-heavy teams often prefer the logical rigor of Fibonacci.
- Integration with tooling: Many agile tools default to Fibonacci, making it the path of least resistance.
Fibonacci works particularly well for teams doing continuous integration and deployment where precise capacity planning matters. When you're committing to specific stories for two-week sprints, the granular scale helps optimize what fits.
When to Choose T-Shirt Sizing
Choose T-shirt sizing when you need:
- Fast initial adoption: Brand new agile teams benefit from T-shirts' immediate comprehensibility.
- Roadmap and release planning: When estimating 100+ items for quarterly planning, sizes work faster than points.
- Mixed-discipline teams: Teams with designers, product managers, and business analysts alongside developers find sizes more inclusive.
- Stakeholder-facing estimation: If estimates get shared directly with executives or customers, sizes communicate more naturally.
- Rough order of magnitude: When precision isn't necessary and you're just bucketing work into categories, T-shirts suffice.
T-shirt sizing excels for product discovery and early roadmapping. When you're evaluating dozens of potential features to prioritize for next quarter, quickly categorizing them as small, medium, or large enables faster strategic decisions than detailed point estimation.
Hybrid Approaches
Many successful teams use both methods for different purposes. They estimate individual user stories with Fibonacci for sprint planning but size epics and features with T-shirts for roadmap planning. This "dual scale" approach leverages each method's strengths.
The conversion path matters. Teams that start with T-shirt sizing during initial backlog grooming often convert to Fibonacci as work approaches active development. This progression mirrors increasing knowledge: rough sizing when information is scarce, precise points when you're ready to commit.
Some teams create custom mappings: XS=1, S=2, M=3, L=5, XL=8, XXL=13—literally Fibonacci with shirt labels. This provides T-shirts' accessibility while maintaining Fibonacci's mathematical properties. The hybrid approach works if applied consistently, though it adds complexity.
Tools supporting both methods, similar to how verification systems adapt to different security contexts, give teams flexibility to experiment. Try Fibonacci for three sprints, switch to T-shirts, measure velocity stability, and choose based on data rather than preference.
Common Pitfalls with Both Methods
Regardless of scale choice, teams fall into similar traps. Treating estimates as commitments rather than predictions destroys psychological safety. Converting either points or sizes to hour estimates undermines the abstraction both methods provide.
Comparing velocity across teams causes dysfunction with any scale. Whether Team A completes 40 points or 2 XL + 3 L stories while Team B completes 25 points or 5 M stories doesn't indicate which team is more productive. Context, skill, and domain complexity vary too much for cross-team comparison.
Letting estimation scales become permanent identity creates resistance to change. "We're a Fibonacci team" or "We're a T-shirt team" becomes dogma. The best teams stay pragmatic: "We use what works and adapt when circumstances change."
Making Your Decision
Run a simple experiment. Take ten representative stories from your backlog. Have the team estimate five with Fibonacci and five with T-shirts. Then ask:
- Which method sparked better discussion?
- Which felt more natural for this team?
- Which scale would stakeholders understand better?
- Which supports our planning process better?
The answers reveal what works for your specific context. Don't choose based on what other teams do or what agile purists recommend. Choose based on what helps your team deliver value predictably.
Both Fibonacci and T-shirt sizing are tools, not religions. The best estimation method is whichever your team applies consistently, reflects on regularly, and improves continuously. Perfect adherence to Fibonacci practiced poorly beats imperfect T-shirt sizing practiced well never happens—good enough consistently beats theoretically optimal inconsistently.
Start with whichever method feels more accessible to your team. Give it 5-6 sprints to stabilize. Track velocity regardless of scale. If you're getting predictable delivery and healthy team dynamics, you've chosen correctly. If not, try the other method. The cost of switching is lower than the cost of persisting with the wrong approach.