About Pink Elephants, Giant Giraffes
summary
Claims that qualitative risk matrices hide wide disagreement and urges replacing labels with 0–100% probabilities and impacts as 90% confidence interval in euros. Shows Monte Carlo runs (Excel or Python) to produce a loss exceedance curve, uses the Brier score for feedback, cites Good Judgment superforecasters and the Chaos Report (project managers correlate with 3x–4x lower success rates), and recounts building a product, buying it back and turning it into a viable business.
AI-written summary · about
chapters (19)
- 0:00Welcome
- 1:25Giraffes and risk beliefs
- 3:33Risk management and chance
- 5:44Cynefin decision contexts
- 8:15Probing complex environments
- 9:41Analysis paralysis
- 11:08Intuition in decisions
- 13:18Probability and risks
- 16:32Risk matrices fail
- 20:26Quantitative risk assessment
- 24:24Probability management
- 27:20Uncertain distributions
- 29:06Monte Carlo simulations
- 31:34Loss exceedance curves
- 32:19Risk mitigation strategies
- 33:24Brier score evaluation
- 35:12Super forecasters
- 36:38Key takeaways
- 38:46Q&A
related talks
11mEstimates as probabilities
34mWhat if? Simulation in portfolio management and replacing estimation as risk strategy
41mCollective Risk Management
35mLeader in the Land of Dilemmas
45mReducing risk in projects, increasing resilience to unexpected change
30mSoftware Estimation Made Easy
51mRespect estimates
53mHow Not to Be Wrong in Software Engineering