Story Complexity Assessment Methods
How do you assess complexity beyond just "how much code to write"? Two stories might be same size but one touches critical auth system and one is isolated. How do you factor this into estimates?
Three-dimension estimation: Effort + Complexity + Uncertainty. Effort = coding time, Complexity = technical difficulty/risk, Uncertainty = how well we understand it. Story A might be 3 effort, 8 complexity, 5 uncertainty = estimate 8. Story B is 5 effort, 2 complexity, 1 uncertainty = estimate 5.
Higher dimension wins. This makes complexity explicit instead of hidden.
Ask: What's the blast radius if this breaks? Touching auth = whole system down = automatic +3 points minimum. Isolated feature = only that feature affected = no complexity premium. Makes risk concrete in estimates.
Count integration points. Story touching 1 system = base estimate. Each additional integration = +2 points. Story touching 4 systems gets automatic +6. Complexity usually comes from integration, not individual pieces.
Code coverage as complexity indicator. <50% test coverage in that area? Add uncertainty buffer. >80% coverage? Reduce estimate - you can refactor safely. Tests don't lie about complexity.