Checkpoint: Choosing Under Uncertainty

Quiz

1. Maya is choosing between a retainer client and a product launch. Which comparison correctly measures the opportunity cost of taking the retainer?
   • The value of the best alternative she gives up by taking the retainer, such as the launch she cannot ship.
   • The total cash the retainer pays, since that is what she gains.
   • The value of the ideal version of the retainer, assuming everything goes perfectly.
   • The value of doing nothing, since staying put costs her nothing.
2. Which of these are table defects described in the lesson? Select all that apply.
   • Score precision theater, such as 4.37 out of 5 implying measurement you do not have.
   • Weight laundering: adjusting weights after seeing which option wins.
   • Blank cost cells that silently read as zero.
   • Dropping a row where every option scores the same.
   • Including the status quo as a scored column.
3. In a comparison table, weights are your preferences about what matters, while scores should be backed by ____ wherever a fact exists.
4. A team says, "We have already put two years into this tool, so we should keep going." Which bias is driving that reasoning?
   • Availability
   • Planning fallacy
   • Sunk cost
   • Anchoring
5. Which practices make a pre-mortem produce useful output rather than a worry list? Select all that apply.
   • For each top failure mode, name the bias that would hide it and one cheap check to run before committing.
   • Run the pre-mortem before the comparison is locked, while checks could still change the outcome.
   • Aim for a long list of abstract worries to show thoroughness.
   • Debate and rebut each failure mode as soon as it is raised.
   • Write failure modes silently before anyone speaks, so the loudest voice does not anchor the list.
6. A decision is a two-way door when its ____ cost is low, meaning it can be undone cheaply and quickly.
7. A six-person team spends three weeks comparing note-taking apps for a trial that can be cancelled monthly. What is the best diagnosis and fix?
   • They need a higher confidence bar, around 85 percent, before trying the app.
   • They over-processed a two-way door; they should time-box the choice and decide with enough evidence to start.
   • They under-processed a one-way door; they should add a pre-mortem and a challenger.
   • They should keep comparing until no uncertainty remains.
8. Take one decision from your project brief. State a confidence range as a number for a specific claim (for example, a date range or a percent chance of an outcome), then explain how you would check whether that range is calibrated.

Sign in to submit answers.