Agile | Scrum | Project Management | Software Engineering | Testing | Continuous Integration Fibonacci Theater: Adding Mathematical Mystique to Scrum Estimation January 18, 2025
Neil Chaudhuri

Neil Chaudhuri

LinkedIn Image of Neil Chaudhuri

I saw this question on LinkedIn and hear it a lot, so I figured I’d write a post for anyone else wondering the same thing.

Scrum Masters & Agile Coaches: Why do we use Fibonacci Sequence 1,2,3,5,8 etc. for estimation instead of just Natural numbers like 1,2,3,4 etc?

Story points are the fundamental metric in Scrum, which through sustained, effective marketing has become the default agile methodology. Story points have been a source of confusion for decades because it’s easy to mistake what they represent. Story points are not hours or a measure of anything particularly concrete but rather a fuzzy representation of the size, scope, complexity, and value of a software feature relative to one another. Assigning one feature 2 points is marginally useful, but knowing that feature is a lot easier than that 8 over there is much more useful.

Another fundamental notion in Scrum, and really agile writ large, is uncertainty, which is a profound element of software engineering. We need to be intentional about recognizing what we don’t know yet. Story point estimates grow less and less useful with more and more uncertainty. Because the Fibonacci Sequence grows exponentially rather than linearly, Fibonacci numbers are a clever (if gimmicky) symbol of an idea—the idea that even the slightest increase in uncertainty yields a significant increase in our story point estimate.

It is really important to remember that the inclusion of fancy-seeming math-iness like Fibonacci in Scrum is not some kind of research-derived correlation between engineering effort and productivity or anything remotely provocative like that. It’s just symbolic of the consequences of uncertainty.

The confusion around Fibonacci’s relevance to Scrum is one example of why I believe story points are the key reason Scrum fails so often, but that is a story for another day.