[Workshop] NoEstimates: Forecasting with Less Effort and More Accuracy
summary
A hands‑on simulation runs 25 story cards with known effort (average 21) and 4 effort points per person per day; teams vary in deterministic estimates (21–28 days) despite identical backlogs, and correlation between estimated effort and actual delivery time is near zero. The game uses dependency cards, a blocker spinner and a deployment die (a roll of 1 forces rework), and shows that tightening work in progress and changing collaboration policy improves predictability. The practical alternative is a probabilistic forecast from completed-item data rather than a single-point no estimates guess.
AI-written summary · about
chapters (9)
- 0:00Welcome and introduction
- 1:16Game setup and rules
- 13:19Team roles and dependencies
- 21:25Initial estimates and variations
- 28:26Debrief and lessons learned
- 38:18Sources of variation in delivery
- 40:06Correlation between effort and time
- 43:13Probabilistic forecasting
- 46:14Summary and resources
related talks
03h 00mBeyond Estimates (Estimates and NoEstimates ) - Let's Explore the Possibilities
41mSo you think you can estimate? The lost art of useful estimates
48mRespect estimates
01h 45mEstimates or NoEstimates
30mSoftware Estimation Made Easy
45mStop rolling the dice: an approach to effort estimation in software development
47mCreating efficient, accurate, software estimations
48mNoEstimates