Advance Decision Markets: Kleros Fellowship Offers Up to $5,000

Build synthetic stable assets from ETH payoff pairs that are options-like instead of debt. This could remove liquidations and allow slower, safer oracles like Kleros. See this post from V. Buterin.

• • Research whether ETH-collateralized synthetic assets can avoid liquidations by using options-like payoff pairs instead of debt.
  • Compare this model to liquidation-based stablecoins/perpetuals on oracle security, MEV risk, capital efficiency, and failure modes.
  • Study user rebalancing, slippage-minimizing market design, and prediction-market-style scalar options as the core mechanism.

Use conditional markets on token/stock prices to decide whether DAO proposals should pass. See this post from Distbit. We can collaborate with futarchy.fi.

• • Research whether asset futarchy can improve DAO governance by using conditional markets on token prices to evaluate proposals.
  • Compare asset-price maximization with KPI-based futarchy, especially for long-term value creation and public-good projects.
  • Study manipulation risks, liquidity requirements, and whether token price reflects stakeholder welfare.
  • Test designs using proposal/pass-fail markets, delayed execution, and anti-manipulation mechanisms.
  • Evaluate when asset futarchy is robust enough to govern treasuries, protocol upgrades, or parameter changes.

Use conditional markets to predict whether a decision would lead to a good KPI. See this post from Distbit. We can run the required experiment on Foresight.

• • Research whether KPI futarchy can improve DAO governance by making proposal execution conditional on predicted changes to explicit performance metrics.
  • Compare KPI choice, oracle design, and metric gaming risks across revenue, user growth, decentralization, security, and public-good outcomes.
  • Study how conditional prediction markets aggregate information about which proposals improve the chosen KPI.
  • Test designs with delayed resolution, multi-KPI scoring, and safeguards against short-term metric manipulation.
  • Evaluate when KPI futarchy outperforms token-price-based governance or simple voting.

Use prediction markets to reward outsiders for discovering valuable opportunities early. Applications include VC deals, hiring, grants, music, and research scouting. See this post from D. White and M. Liston. 

• • Research whether private prediction markets can turn opportunity discovery into a scalable, incentive-compatible scouting mechanism.
  • Study sponsors such as VCs, record labels, grantmakers, employers, and research labs that benefit from early signals.
  • Analyze privacy windows, hidden prices, liquidity subsidies, and delayed revelation as tools against free-riding.
  • Compare AMM, order book, and TEE-based designs for private opportunity markets.
  • Evaluate risks of sponsor self-dealing, manipulation, and whether reputation is enough to sustain trader trust. 

Create conditional DeFi “worlds” where assets, lending, AMMs, and derivatives exist only if a future event happens. This enables decision-contingent finance and hedging. See this post from D. White.

• • Research whether “multiverse finance” can reuse prediction market logic to make full conditional financial systems.
  • Study how assets, lending, AMMs, and derivatives can safely compose inside event-specific “verses.”
  • Compare the results with ordinary conditional tokens, futarchy, and options-based synthetics.
  • Analyze oracle resolution, liquidity fragmentation, UX complexity, and contagion between conditional worlds.
  • Evaluate whether multiverse-aware DeFi enables better hedging, policy markets, and decision-contingent capital allocation.
  • Study how the parent universe can interact with child universes.

Replace scalar markets with full probability distribution markets. Useful for forecasting dates, prices, KPIs, grant impact, and tail risks. See this article from D. White. 

• • Research whether distribution markets can improve scalar prediction markets by eliciting full probability distributions instead of bucketed outcomes.
  • Apply them to futarchy, where proposals are compared by the predicted distribution of KPI or token-price outcomes.
  • Study whether decision rules should optimize expected value, downside risk, or tail-adjusted utility.
  • Compare L2-function AMMs with LMSR, LS-LMSR, and bucketed AMMs for capital efficiency and manipulation resistance.
  • Evaluate use cases in DAO governance, grant allocation, attendance forecasting, and risk-aware treasury management.

• Bundle many rare events into one capital-efficient market. Such markets can price crypto tail risks like hacks, slashing, bridge failures, oracle failures, or stablecoin collapses. See this article from V. Buterin.
• • Research whether prediction markets can price many rare tail risks more efficiently by bundling them into a single market.
  • Compare bundled rare-event markets with independent binary markets, LMSR, multiscalar markets, and insurance-pool designs.
  • Study how event correlation, bundle construction, and partial resolution affect pricing accuracy and trader understanding.
  • Apply the mechanism to crypto tail risks such as stablecoin failures, rollup hacks, slashing, bridge failures, oracle failures, and lending-market crashes.
  • Evaluate whether these markets can become useful risk dashboards for protocols, insurers, DAOs, and treasury managers.

Let AI agents trade, evaluate, and help resolve thousands of tiny prediction markets. Use cases include scam detection, AI evaluation, flight delay prediction, and recommender systems. See “AI as a player in a game” in this post from V. Buterin.

• • AI micromarkets for info defense: Research whether tiny prediction markets can classify scams, fake accounts, malicious dapps, token authenticity, spam, or policy violations better than centralized moderation or wallet heuristics.
  • Market-based AI evaluation / decentralized RLHF: Design a system where AI agents earn or lose money across many small prediction tasks, creating market incentives to build generally accurate agents rather than one-off predictors.
  • AI + Kleros-style escalation: Study multi-round resolution where AI agents handle cheap first-pass adjudication, and humans only intervene when enough money accumulates on both sides of a dispute.
  • Closed agents, open rules: Vitalik’s key security point is that individual AIs can remain closed, while the mechanism is open; research whether this reduces adversarial ML attacks compared with open-source AIs.
  • Micromarket liquidity subsidies: Experiment with how little subsidy is needed for AI agents to produce useful prices on small markets: $1, $5, $50, etc.
  • AI agents as market makers for long-tail markets: Study whether AI agents can create, seed, trade, and resolve thousands of low-volume markets that would be uneconomical for humans.

Legal and technical framework for Kleros to serve as a dispute resolution layer for prediction market platforms operating under regulatory oversight (CFTC, MiCA, etc.)

Analysis of how centralized and decentralized oracles fail in practice (manipulation, ambiguity, downtime) and how Kleros’ mechanism design addresses each failure mode

Protocol-level proposals for integrating Kleros as an optional or default oracle in platforms like Polymarket, Kalshi or other Prediction Markets, including comparisons with existing solutions such as UMA’s Optimistic Oracle

How Kleros’ transparent, auditable, and appellable process maps onto requirements from financial regulators for designated contract market (DCM) operators

Empirical analysis of real market resolution controversies and how Kleros’ dispute resolution mechanism would have produced different outcomes