Artifact: An Hour Estimate vs. a Relative Estimate

ex-34 · exercises co-19 · the same task, estimated two ways.

Task: "add multi-currency support to the gift-card redemption flow."

Estimate 1 -- raw hours: "14 hours."

Estimate 2 -- relative, against a reference story: "about the same size as add-partial-redemption (a completed story, 5 points), maybe slightly larger because of the currency-conversion-rate dependency -- call it 8 points."

Verify: the raw-hour estimate ("14 hours") implies a precision the team has no actual basis for -- multi-currency work has never been done in this codebase before, so there is no historical data to calibrate 14 hours against. The relative estimate anchors against a REAL, completed, comparably- sized story, making its uncertainty visible (points, not hours) instead of hidden behind a false- precise number.

Key takeaway: "14 hours" and "8 points, roughly like that other story" convey very different amounts of actual confidence, even though both are single numbers -- one pretends to a precision that does not exist, the other is honest about the comparison it is actually built on.

Why It Matters: a raw-hour estimate for genuinely novel work (no prior story to calibrate against) gets treated as a commitment by everyone downstream (a sales date, a dependent team's own plan) even though it was never more than a guess -- relative estimation against a known reference at least makes the guess's basis visible and checkable.