Week 5 Problem Session: Value of information workshop

Rebuilding the lecture's blind-vs-informed EV comparison, then building a second toy example from scratch with new numbers

22 March 2027

Before the session#

Read the lecture page. Bring the $10 blind / $40 informed / $30 value-of-information numbers — the session checks them by hand before moving on.

In the session#

The group recomputes the lecture’s toy example from the raw payoff table, confirming +$10 blind and $40 informed before anyone is told the answer. Then, working in pairs, everyone builds a second toy payoff table with different numbers and different state probabilities, computes its value of information, and swaps with another pair to check the arithmetic.

Closing discussion: why is the value of information always ≥ 0? Groups argue it from the payoff table’s structure directly, not from having seen the lecture’s result stated as a fact.

Afterwards#

Nothing is marked this week. Going into week 6: comfort computing value of information for an unseen payoff table, since week 6 immediately extends this into partial (not perfect) signals.