TCRs -

A Game Theoretic Approach

Aditya Asgaonkar

Ethereum Foundation (current)

Univ. of Southern Calif. (former)

Acknowledgement

Work done as part of thesis under Prof. Bhaskar Krishnamachari at USC Viterbi

Mechanism Design & Blockchain

  • Engineering economic mechanisms towards desirable objectives
  • Especially relevant in the blockchain space, since we have payoff functions readily available to us
    • Staking rewards
    • Token values

Token Curated Registries

  • Registry: A list that contains information of value to the public
  • Token: An associated “share” in this registry
  • Curation: The registry is maintained by a set of participants - the token holders

adChain

Kleros T2CR

TCR Incentive Model

  • Objective: Perfectly curated list of items

  • Assumption: Token value is linked to the quality of the registry

TCR Incentive Model

  • Application process:
    • Candidates apply to the registry by staking some tokens
    • Token holders vote on whether to accept or reject the candidate from the registry
    • If candidate is rejected, then their staked tokens are slashed and distributed to the winning voters
    • The winning side of the vote takes some fraction of the token from the losing side

TCR Incentive Model

  • Good candidates are accepted
    • Quality of TCR improves
    • Token value increases
  • Bad candidates are rejected
    • Quality of TCR doesn’t degrade
    • Token value is maintained
    • Candidate’s staked tokens are distributed among winning voters

TCR Incentive Model

  • Possible outcomes:
    • Candidate accepted
      • Same token amounts
      • Increased token value
    • Candidate rejected
      • Increased token amount
      • Same token value

TCR Incentive Model Analysis

  • (Assumption) Token price is directly correlated to TCR quality

TCR Incentive Model Analysis

  • Good candidates will improve TCR quality

TCR Incentive Model Analysis

  • Bad candidates will degrade TCR quality

TCR Incentive Model Analysis

  • What happens when candidate is only marginally good?

TCR Incentive Model Analysis

  • What happens when candidate is only marginally good?
  • Dilemma for voters:
    • Accept: Slightly higher valued tokens
    • Reject: More tokens from slashed candidate

TCR Incentive Model Analysis

  • What happens when candidate is only marginally good?
  • Dilemma for voters:
    • Accept: Slightly higher valued tokens
    • Reject: More tokens from slashed candidate

??

TCR Incentive Model Analysis

  • For the same set of candidates, the order of application matters!

Well, actually…

This problem arises because of distributing slashed tokens!

  • Why is distributing slashed tokens necessary?
  • Isn’t just token price sufficient?
  • Why is distributing slashed tokens necessary?
  • Isn’t just token price sufficient?

It isn’t!

It is!

Why have slashed token distribution in the first place?

Why have slashed token distribution in the first place?

To incentivize holders to vote

Just having a token is sufficient...

  • No token amount vs token value dilemma!
  • Natural incentive to vote; inactivity may drive token price down!

Conclusion

  • Sound mechanism design is essential to the security of your dapp!
  • Complicated mechanisms are hard to analyze!
  • Seemingly simple modifications to mechanisms have unusual side-effects!

TCRs - A Game Theoretic Approach

Thank You!

Twitter: @adiasg

TCRs - A Game Theoretic Approach - Google Slides