Assignment Overview: This case study assignment asks students to analyze a real-world cryptoeconomic failure using game theory and mechanism design concepts from L07. Unlike A07 (Audit Challenge), which focuses on code-level vulnerabilities, A11 focuses on system-level mechanism design flaws. The two assignments complement each other: A07 examines micro-level smart contract bugs; A11 examines macro-level economic design failures.

Learning Objectives & Course Alignment

Core Learning Goals:
  1. Analytical Thinking: Apply formal game theory concepts to real events (not just textbook examples)
  2. Mechanism Design Critique: Identify specific design flaws and articulate why they led to failure
  3. Constructive Reasoning: Propose improvements with honest trade-off analysis
  4. Communication: Distill complex system failures into clear, time-limited presentations

Course Integration

This assignment works best after students have covered:

Optionally, A07 (Audit Challenge) provides useful context for Option C (The DAO hack), as students will already understand reentrancy.

Pre-Class Preparation

Time Required

Activity Instructor Time Class Time
Pre-class preparation 15-20 minutes -
Introduction & event overview - 5 minutes
Phase 1: Timeline reconstruction - 15 minutes
Phase 2: Root cause analysis - 15 minutes
Phase 3: Mechanism improvements - 10 minutes
Presentations & discussion - 5 minutes
Total Class Time - 50 minutes

Materials Checklist

For Each Student/Pair:
  • Access to instructions page (digital or printed)
  • Access to course slides for L07 (for reference during analysis)
Instructor Materials:
  • Answer key (for all 3 event options)
  • Grading rubric
  • Timer for presentations
  • Whiteboard/projector for wrap-up discussion

Event Assignment Strategy

Recommendation: Let students choose freely, but aim for a mix across all three events. If too many students cluster on one event:
  • Terra/Luna tends to be most popular (most recent, most DeFi-relevant)
  • FTX generates the most debate (fraud vs. mechanism failure)
  • The DAO connects best to A07 (Audit Challenge) and is good for students who completed that assignment
If all students choose the same event, consider assigning events to ensure diversity in presentations.

Minute-by-Minute Teaching Plan

Minutes 0-5: Introduction

Opening Hook (2 minutes):

"Between 2016 and 2022, three crypto failures destroyed over $70 billion in value. The DAO lost $60 million from a single code bug. Terra/Luna vaporized $40 billion in a week from a flawed economic mechanism. FTX destroyed $32 billion from old-fashioned fraud enabled by lack of transparency. Today, you're going to figure out why -- not just what happened, but what mechanism design failures made these collapses inevitable."

Activity Explanation (3 minutes):
  • Emphasize: this is analysis, not just retelling the story. "I want to see Nash equilibrium, dominant strategies, incentive compatibility -- the concepts from L07."
  • Clarify the 4 phases and time limits
  • Note that all 3 events are equally valid choices
  • Remind students: 2 mechanism improvements, not just 1

Minutes 5-20: Phase 1 -- Timeline Reconstruction

Instructor Role:
  • Circulate and check that students are not just copying the key dates from the instructions. They should add their own actor/incentive analysis
  • Prompt with: "Who are the key actors at this point? What are their options? What's their dominant strategy?"
  • If students struggle to identify game dynamics, suggest: "Think about what happens if everyone acts rationally at this moment. Does that lead to a good or bad outcome for the system?"

Minutes 20-35: Phase 2 -- Root Cause Analysis

This is the most important phase. Push students beyond narrative description.
  • For Terra/Luna students: "Can you draw the feedback loop? UST selling leads to... which leads to... which leads to more UST selling. Is this incentive-compatible?"
  • For FTX students: "Was this a mechanism design failure or just fraud? Can you design a mechanism that prevents even a dishonest exchange operator from stealing funds?"
  • For DAO students: "The vulnerability was known 3 weeks before the attack. Why wasn't it fixed? What does that tell you about governance mechanism design?"
Common Struggles & Interventions:
Struggle Intervention
"I'm just describing what happened" "Stop narrating. Draw a payoff matrix for the key decision point. What are the players' options? What's the Nash equilibrium?"
"I can't find the mechanism design flaw" "What assumption did the system designers make about how people would behave? Was that assumption correct?"
"My improvements are vague" "Make it specific enough to implement. 'More transparency' becomes 'publish Merkle-tree proof of reserves every 24 hours.'"
"FTX was just fraud, not mechanism design" "Great insight -- but what mechanism could have prevented the fraud? That's the design challenge."

Minutes 35-45: Phase 3 -- Mechanism Improvements

Key Guidance:
  • Insist on exactly 2 proposals. Students often want to list 5 vague ideas instead of developing 2 deeply
  • Each proposal must include a trade-off. "What's the cost of your improvement? What do you give up?"
  • Reference L07 concepts explicitly. "What L07 concept does your proposal implement? Is it a commitment device? A screening mechanism?"

Minutes 45-50: Presentations & Discussion

Presentation Management:
  • Select 1-2 students/pairs to present (aim for different events if possible)
  • Strict 5-minute timer per presentation
  • Encourage questions from classmates analyzing different events
Wrap-Up Discussion Prompts (if time permits):
  • "What common pattern do you see across all three failures?" (Answer: reflexivity, information asymmetry, speed of collapse)
  • "Could any of these failures have been prevented by better regulation alone?" (Opens debate about mechanism design vs. regulation)
  • "Which of the three events was most clearly a mechanism design failure vs. a human failure?"
  • "Do you think 'code is law' is a viable principle after studying these events?"

Discussion Facilitation Tips

Socratic Questions by Event

For Terra/Luna:
  • "If you held UST on May 8, what would you do? Why?" (Reveals bank run incentive)
  • "Could the peg have been saved if everyone just held? Is 'everyone holds' a Nash equilibrium?" (No -- selling is dominant)
  • "What's the difference between Terra/Luna and a regular bank run?" (Reflexive token relationship makes it worse)
For FTX:
  • "How is FTX different from a bank?" (Banks have deposit insurance, regulatory oversight, reserve requirements)
  • "If you were designing a trustless exchange, what would you change?" (Leads to proof-of-reserves, on-chain custody)
  • "Was CZ's tweet a strategic move or responsible disclosure?" (Introduces signaling games)
For The DAO:
  • "The vulnerability was public knowledge for 3 weeks. Why wasn't it fixed?" (Governance speed vs. attack speed)
  • "Was the hard fork the right decision? What does it mean for immutability?" (No right answer -- tests reasoning quality)
  • "How did the 28-day withdrawal delay save the day? Was that intentional?" (Accidental commitment device)

Common Misconceptions to Address

Misconception Reality
"Algorithmic stablecoins can't work" Terra's specific design was flawed (reflexive), but other algorithmic approaches (e.g., over-collateralized like DAI) have survived stress tests. The issue is the mechanism, not the category.
"FTX failed because crypto is a scam" FTX failed because of fraud and lack of transparency -- problems that exist in traditional finance too (Enron, Wirecard). The crypto-specific lesson is about designing systems that are trustless.
"The DAO hard fork proved blockchain is centralized" The fork required community consensus and created a permanent split (ETC). It demonstrates that governance is inescapable, not that decentralization is impossible.
"Better regulation would have prevented all of these" Regulation could help (especially FTX), but Terra's failure was a mechanism design problem that regulation might not have caught. The DAO's failure was a code-level bug. Different failures need different solutions.

Assessment & Grading Tips

Quick Grading Strategy

During Presentations (take notes on):

  • Does the student use L07 vocabulary correctly? (Nash equilibrium, incentive compatibility, etc.)
  • Is the root cause analysis specific or generic?
  • Are improvement proposals concrete with trade-offs?

Written Work Grading (post-class):

  • Batch by event: Grade all Terra/Luna analyses first, then FTX, then DAO. Maintains consistency within event.
  • Focus on reasoning quality: A student who correctly argues "FTX was fraud, not mechanism design" deserves high marks for critical thinking -- as long as they still propose mechanisms that would prevent fraud.
  • Time estimate: 3-5 minutes per student/pair at the 10-point scale

Recognizing Excellence

Look for and reward:
  • Formal payoff analysis: Students who draw actual payoff matrices or game trees
  • Novel mechanism proposals: Ideas beyond the obvious (e.g., "insurance pool funded by tx fees" rather than just "more collateral")
  • Cross-event connections: "Terra and FTX both involved reflexive collateral -- using your own token as backing"
  • Honest trade-off analysis: Students who identify weaknesses in their own proposals
  • Philosophical depth: Engaging with "code is law" vs. governance flexibility (especially for DAO students)

Adaptations for Different Class Sizes

Class Size Modifications
Small (10-15) Individual work. All students present (3 minutes each). More discussion time. Assign events to ensure all 3 are covered.
Medium (20-30) Pairs. Select 3 presenters (one per event). Standard 50-minute format.
Large (40+) Groups of 3. Select 3-4 presenters. Consider extending to 60 minutes for discussion. Alternatively, have groups prepare written submissions only and select best for next-class discussion.

Extension Activities

For Advanced Students

  • Formal game model: Create an explicit payoff matrix or extensive-form game tree for a key decision point
  • Comparative analysis: Analyze all 3 events and identify common mechanism design patterns
  • Research paper: Extend analysis into a 5-page paper with academic references
  • Simulation: Build a simple Python model of the Terra death spiral showing how parameters affect stability

For Struggling Students

  • Pre-class reading: Assign a news summary article about each event as homework
  • Concept cheat sheet: Provide a one-page summary of L07 concepts with definitions
  • Scaffolded analysis: Provide a template: "The mechanism design flaw was ___. This violated incentive compatibility because ___. A Nash equilibrium analysis shows ___."
  • Focus on one phase: Allow struggling students to do a deep timeline + root cause analysis without the mechanism improvement phase

Follow-Up Opportunities

Next Class Session:
  • Show short video clips of key moments (CZ tweet, Do Kwon interview, SBF testimony)
  • Compare student proposals with actual post-event reforms (e.g., proof-of-reserves adoption after FTX)
  • Discuss how the crypto industry has (or hasn't) learned from these failures
Connection to Other Assignments:
  • A07 (Audit Challenge): The DAO hack connects directly -- students who did A07 will recognize reentrancy
  • A05 (Token Design): Terra/Luna connects to token economics -- what happens when tokenomics are reflexive?
  • A09 (DAO Proposal): The DAO hack raises questions about governance design relevant to A09

Troubleshooting Common Issues

Problem Solution
Students just retell the story Interrupt early: "I can read the Wikipedia article too. What's the Nash equilibrium? What's the dominant strategy? That's what I want to see."
Everyone chooses the same event Assign events to balance. Or let it happen and use presentations to compare different perspectives on the same event.
Students lack background on events Spend 5 extra minutes on the opening overview. Provide optional pre-reading links.
Mechanism improvements are too vague Apply the "could you build this?" test: if a developer couldn't implement the proposal, it's too vague.
"This was just fraud/bad people" Agree, then redirect: "Yes, but what mechanism would have prevented a bad actor from succeeding? That's the design challenge."
Students struggle with game theory vocabulary Refer them to L07 slides. Provide a quick 2-minute refresher on Nash equilibrium and dominant strategies at the start.
Important Reminder: This assignment asks students to analyze real events that caused real financial harm to real people. Encourage analytical rigor but also sensitivity. Some students may have personal connections to these events (e.g., lost money in Terra or FTX). Frame the discussion as "what can we learn?" rather than "who was stupid?"

© Joerg Osterrieder 2025-2026. All rights reserved.