Another TWST3 write-up on another presentation that I didn’t write down the title of (and yes, I know this was 6 weeks ago). From my spotty memory and notes however, it was a general talk on doing risk analysis for software projects. The hand-wavey nature of the description might make this less useful for anyone other than me, but hey, it’s my blog. 🙂

  • Probability and Impact are opinions, and are therefore influenced by the opinion holder’s biases
  • Probability and Impact numbers might also reflect a person’s negotiating skill
  • Know your own personal strengths and weaknesses around your risk analysis
  • The part of their presentation that got the most discussion was their use of affinity analysis which I don’t have a handy link to a nice describing page
  • This talk was also the catalyst for the Magic numbers discussion that was almost guaranteed to happen. I have a bunch of chicken scratch with no context written down, but the most clear point is ‘How much bigger is a medium than a high?’
  • Instead of using buckets based upon some number system, use labels such as intolerable
  • Challenge the value of a number
  • Challenge the source of the number
  • A system that was attributed to Tom Gibb was mentioned for visually clarifying the quality of a number
    • Bold has high credibility
    • Italics has low credibility
    • Regular has neutral credibiltiy
  • Statistics categorize things the way they were organized
  • How good / complete is the dataset you are going to use to generate the risk numbers?
  • What is Risk?
    1. A Bad Thing could happen
    2. Question asked in Parliament (a knock-on effect of the first)
    3. A willing gamble
  • The person who takes the meeting minutes makes the history
  • Bad risk analysis is a self-correcting problem – James Bach
  • Send a confirmation email about your understanding of the events of a meeting — or better yet, use RSS or similar asynchronous channel