Obviousness Around the Clock
summary
Replicates and extends Changu’s experiment: groups of four, 20 auctions, affiliated private values (group draw 10–110 plus individual 0–20), ~70 subjects per treatment, average payout €24. Compares sealed‑bid second‑price, ascending clock, and two hybrids that show an auction clock passively after bids (one passive clock without counts, one with live dropout information). Finds most of AC’s advantage comes from feedback: showing the clock dramatically cuts mean absolute deviations from truthful bids within 2–3 rounds; revealing live dropout counts reduces underbidding further.
AI-written summary · about
chapters (13)
- 0:00Intro
- 0:27Background
- 5:50Obvious Dominance and Strategy Proofness
- 9:28Obviousness and Evidence
- 11:32Decomposing Obviousness: 2P vs. AC Auctions
- 17:48Experimental Design
- 21:12Experimental Results: Mean Absolute Deviations
- 26:59Discussion: Behavior and Treatments
- 28:13Distributions of "Bids minus Values"
- 31:57Approach
- 36:01Analysis of Incentives
- 38:14Discussion: Behavior and Incentives
- 40:28Final Remarks
related talks
46mFrom correlation to causation: measuring ad effectiveness at scale
47mDebate "The force behind decision-making – data, wisdom, and the crowds"
21mLeveraging Multi-Armed Bandit Algorithms for Dynamic Decision Making
26mThe specifics of the AB-test for Unit Economics with a network effect
20mThe force behind decision-making – data, wisdom, and the crowds
15mBeyond Efficiency
35mLearning to Judge
30mOptimizing Mobile Game LiveOps with Data-Driven Decisions