The final round, simulated

Assessment Day

Three stations. Every decision graded.

  1. The Betting Table, 8 priced propositions. Pass or bet with a size. Graded against the true edge and the Kelly fraction: betting −EV is fatal, oversizing +EV nearly so.
  2. The Jar Auction, 3 sealed-bid auctions for a jar of uncertain value against three rivals. You see a noisy signal; winning means you probably saw the highest one. Graded against the optimal bid shading.
  3. Calibration, 8 quantities, and you give a 90% confidence interval for each. Graded on whether “90%” means 90% from your mouth, then on tightness.
The rule of the room: we never grade results, a won bet can be a terrible decision and a lost auction a perfect one. The report at the end reads like the hiring committee's notes, because that's how they actually think.