Why relative beats absolute
Mike Cohn's argument (Agile Estimating and Planning, 2005): humans are bad at predicting time but reasonably good at comparing sizes. Asking "is this bigger or smaller than that other story we shipped?" gets you a usable estimate faster than asking "how many hours."
Story points (Fibonacci-ish)
Common scale: 1, 2, 3, 5, 8, 13, 20, 40, 100. The gaps widen on purpose — a story isn't a 7, it's either a 5 or an 8. Forcing the choice keeps the conversation honest. A 13 is a warning sign (likely needs to be split). A 40 means you don't understand it yet.
T-shirt sizing
XS / S / M / L / XL — even less precise than points, but useful for rough early-stage triage where points would imply a confidence you don't have. Common in roadmap prioritization before any specific story exists.
Planning poker
The team estimates each story simultaneously, revealing numbers at the same time. If estimates diverge, the high and low estimators explain — usually the discussion surfaces a hidden assumption. The technique attacks anchoring bias(the first number stated influencing everyone else).
↳ the points-as-time mistake