Lecture 31 — Strategy Under Uncertainty

Position in the course

This is lecture 31 of 32. It belongs to the course arc described in the course outline and uses the standard interactive case method: display, attempt (5 minutes), discuss, reveal, name the technique.

Learning objectives

By the end of this lecture students can do all three under time pressure, which is what the four cases exercise.

Prerequisite knowledge

Before this lecture: Lectures 26 and 29 — probability and simulation

Assumed fluencies:

Preparation task: Compute the expected value of a 10-ticket lottery with published odds.

Detailed topic outline

Core (mandatory):

  1. expected value as a decision tool
  2. payoff tables
  3. the switch question resolved by listing
  4. risk aversion beyond expectation

Extension (optional enrichment):

  1. dominance arguments
  2. mechanism thinking: how rules shape behaviour

Case sequence

CaseTitleDifficulty
cs-121The Umbrella DecisionBeginner
cs-122The Game Show SwitchFoundational
cs-123The Insurance RiddleIntermediate
cs-124The Auction BluffAdvanced

Display each case full-screen from the student page (cases cs-121–cs-124); instructor pages cs-121–cs-124 under case-studies/instructor/ hold the full solutions, reveal guides, and discussion prompts.

Timing plan (120 minutes)

MinutesSegment
0–10Framing: where this fits; recall of the previous lecture
10–30Case 1 (cs-121): 5-minute attempt, then discussion and reveal
30–50Case 2 (cs-122): same rhythm
50–70Case 3 (cs-123): same rhythm, harder case
70–75Break
75–100Case 4 (cs-124): same rhythm, hardest case
100–115Synthesis: name the techniques; misconceptions check
115–120Exit questions, homework, next-lecture preview

Timing flexes with the room: if a case ignites the class, let the discussion run and compress the synthesis; if the attempts stall, harvest partial attempts rather than extending the clock.

Worked examples

Instructor demonstration to open the lecture: Umbrella decision (cs-121): the payoff table drawn once, expected values computed twice — with rain 30% and 70%.

Classroom activities

#ActivityMinutesFormat and prompt
1Payoff table gallery8Four decision scenarios turned into payoff tables; the class tours and votes on best actions.
2Switch simulation7cs-122 played 30 times by pairs with forced switch vs forced stay; results pooled.
3Risk interview5Same expected value, different variances: which do you choose? Answers are discussed, not graded.

Board plan

Instructor explanation notes

Reading the room. Expected value is a decision tool, not a gambling lesson: the payoff-table discipline applies to project choices and policy. Risk attitudes are discussed, never mocked.

Running the reveal. cs-122's switch question is resolved by listing, the course's signature move, one last time.

Lecture focus. One sentence to repeat verbatim: Build the payoff table; let expectation argue, then judge risk.

Common misconceptions

Each case's instructor page lists the specific wrong turn to expect; the reveal should address it by name.

Student practice exercises

Core:

  1. Build the umbrella payoff table (cs-121) for both rain probabilities and pick the action.
  2. List the six outcomes of the switch question (cs-122) and compute both strategies' expectations.
  3. Argue the insurance riddle (cs-123) from both sides: expected value says no, risk says yes.

Enrichment (optional):

Formative assessment

Exit questions (choose one; answer on a card in 60 seconds):

  1. What does expected value compare, and what does it ignore? Expected: average outcomes; it ignores variance (risk) unless you add it
  1. Why does switching win in cs-122? Expected: the host's reveal leaks information; 2/3 of outcomes favour switching
  1. When is a dominated strategy visible instantly? Expected: one option's payoffs are worse in every column of the table

Cards are skimmed before the next lecture; three answers become the framing. Persistent wrong answers route students to the review stations of lecture 16 or to office hours.

Homework and preparation

Learning-outcome mapping

Differentiation