Estimation as Uncertainty Reduction
summary
Estimation is uncertainty reduction: replace single-value guesses with range estimates that expose assumptions, calibrate confidence with an equivalent bet, decompose unknowns into nominal/ordinal/interval/rational scales, and model outcomes with Monte Carlo simulations or simple measurements. A random sample of five gives >90% probability the median falls between min and max, so tiny observations often cut entropy a lot. Built examples include an electrical-contractor Quattro Pro estimator and separation of Q‑time from work estimates.
AI-written summary · about
chapters (33)
- 0:00Intro
- 0:23What is this estimation thing, anyway?
- 2:08Precise plans hide uncertainty
- 3:17Leveraging Uncertainty
- 3:56What is this Estimation Thing?
- 4:39The Domain Expert
- 5:17Bundled Assumptions
- 6:05Unbundling Assumptions with Ranges
- 6:44Expressing Uncertainty
- 7:40Reducing Uncertainty
- 8:59Quantifying Uncertainty Reduction
- 10:00A Mathematical Theory of Communication
- 10:44Estimation as Uncertainty Reduction
- 11:24A Definition of Information Entropy
- 13:04Reducing Information Entropy ..how much does 5 random values reduce entropy?
- 13:1690% probability better than "no clue"
- 14:03Confidence Interval
- 15:45A Taxonomy of Measurement Methods
- 16:20Rational Scales: Time and Money
- 16:53Interval Scales: Story Points
- 17:33Mohs Scale of Mineral Hardness
- 18:00Ordinal Scales - 5 Star Rating
- 18:23Conflating ordinal and rational scales
- 18:44Ordinal Scales: The Cockburn Scale
- 19:43Nominal Scales INVES
- 20:06Nominal Scales ...qualitative measurement...
- 21:02Nominal Scales qualitative measurement...
- 22:43Diving Deeper...
- 23:09Estimate Validation Techniques
- 25:08Calibration
- 26:02Story Decomposition ...right-sizing stories is half the battle.
- 26:50Release Planning and Queue Time
- 28:07Monte Carlo Simulations for Project Estimation
related talks
29mSoftware estimation – embracing nuance and controlled chaos
01h 00mEstimation: The Art of the SWAG
47mCreating efficient, accurate, software estimations
01h 45mEstimates or NoEstimates
07mEstimates are BS
42mBeyond Estimates (Estimates and "NoEstimates") - Let's Explore the Possibilities
40mEstimations: from guessing to weighing in grams
30mSoftware Estimation Made Easy
