Software Estimation Made Easy
summary
Argues that expert guesses and story points fail and proposes statistical forecasting based on lead time and WIP distributions, using historical start/end dates to run 10,000 simulations that produce a histogram and give 50% and 90% confidence dates. Describes a free tool called Projector that implements this Monte Carlo approach, counts throughput instead of sizing, and reports that forecasts are cheaper and usually better though not perfect — claimed 90% accuracy often behaves like ~75% because of scope change, team stability and client collaboration.
AI-written summary · about
chapters (6)
- 0:00Start
- 0:29LUNAR LOGIC
- 11:48EXPERT GUESS DOESN'T WORK
- 12:40ALTERNATIVE?
- 13:47A FORECAST FOR JAREK
- 22:18A NEW PROJECT VERSUS AN ONGOING PROJECT
related talks
47mCreating efficient, accurate, software estimations
57mThe Science of not estimating
44mAgile metrics for predicting the future
01h 00mEstimation: The Art of the SWAG
41mYour Project Behaves Like a Hurricane--Forecast It Like One
29mSoftware estimation – embracing nuance and controlled chaos
46mEstimates and complexity: a tale in 42 and a half minutes
42mEstimates: Is There a Silver Bullet?