Game Theory and Blockchain: Nash Equilibrium Meets Crypto
Book: Cryptoeconomics
Authors: Jian Gong, Wei Xu
ISBN: 978-0-367-42993-5
Previous: Optimized Consensus | Next: Behavioral Economics
Chapter 4 is where Gong connects the dots. Consensus is not just computer science. It is mechanism design: set the rules so selfish players produce a good outcome anyway.
Game theory basics
Players. Strategies. Payoffs. Everyone maximizes their own interest and guesses what others will do.
Gong runs through classifications: cooperative vs non-cooperative, static vs dynamic, complete vs incomplete information. The prisoner’s dilemma is the star example.
Two suspects, separate rooms. If both stay silent, light sentence. If one betrays, the betrayer walks. Classic result: both betray even though cooperation is better together.
On a blockchain with no trust, betrayal can pay even more (fork the chain, double-spend). So you need penalties baked into the protocol.
Add a -6 point punishment for betrayal and the Nash equilibrium flips back to cooperation. That is the cryptoeconomic move in one sentence.
Schelling points
Thomas Schelling studied how people coordinate without talking. Ask strangers to pick a number from a list. Many pick 100000000000 because it stands out.
Blockchains need the same magic. Which chain is “the real Bitcoin”? The one with the most work, the longest chain, the one everyone expects everyone else to follow.
Casper later forces validators to bet on what they think the majority will bet. Schelling logic all the way down.
Bounded rationality
Herbert Simon argued humans are not perfect optimizers. We satisfice. We follow habits. Xiao Li does not grab 5 yuan from the gym desk because it breaks his routine, not because he ran a cost-benefit analysis.
Protocol designers who assume hyper-rational actors will get surprised. Real users cut corners, chase hype, and panic sell.
Mechanism design = reverse game theory
Start with the outcome you want. Build rules backward.
Auction design: make truthful bidding the dominant strategy. Consensus design: make honest mining or staking the dominant strategy.
Bitcoin’s PoW is Nakamoto applying that template. He wanted censorship resistance and agreement on ledger state. He assumed miners chase profit. PoW plus rewards plus attack cost delivered the rest.
Bribery and the P + ε attack
This section is the chapter’s payoff.
Imagine a vote on chain state. Vote with the majority, earn P. Sounds stable.
Now an attacker promises: vote minority, get P + ε extra if you are alone on the losing side. Rational players pile onto the minority. Attacker wins without paying ε because the minority never forms.
That is the P + ε attack. Cheap to threaten, expensive to defend against in naive PoW voting.
Fix: require staked deposits. Vote wrong, lose everything. Briber must cover your slash plus reward. Cost explodes.
Casper preview
Gong closes with Casper’s betting game. Validators stake ETH on block outcomes. Flip-flop bets get slashed. Convergence toward one chain is the Schelling coin game with real money on the line.
Threat model summary:
- Strong against uncoordinated bribery (slash makes it hurt)
- Still vulnerable to coordinated 51% style attacks, but at huge capital cost
What I think
Chapter 4 is the book’s theoretical spine. If you read one technical chapter beyond the intro, make it this one.
The payoff matrices look dry. But they explain DAO governance fights, why slashing exists, and why “just vote” systems fail.
Gong wrote this before Ethereum shipped full PoS. Casper details changed. The bribery math did not.