Record estimate before and actual time after every 30+…

When This Fires

• When starting any estimation calibration effort
• When you consistently miss deadlines or underestimate how long things take
• When building the data foundation for Apply your personal estimation ratio as a multiplier… (personal estimation multiplier)
• When you suspect your time estimates are biased but don't have data to confirm

Common Failure Mode

Estimation without feedback: making time estimates but never recording actuals, so you never learn how biased your estimates are. Each estimate is made from scratch, with the same biases, producing the same systematic error. Without the feedback loop, estimation skill never improves.

The Protocol

(1) For every task estimated to take 30+ minutes, write down your estimate before starting. Use whatever capture tool is convenient: a notebook, a spreadsheet, a note on the task itself. (2) After completing the task, immediately record the actual elapsed time. Don't wait — memory of duration degrades quickly. (3) Continue for at least two weeks (or 20+ tasks, whichever comes first) to build a statistically useful sample. (4) Calculate your personal estimation ratio: average(actual time / estimated time) across all tasks. This is your calibration number. (5) Use this ratio going forward as a multiplier (Apply your personal estimation ratio as a multiplier…). Continue tracking to verify the ratio holds and to detect if it changes as your estimation skill improves.