Makes the case that prediction markets are hitting the same infrastructure wall DeFi faced in 2020, with the oracle layer as the bottleneck to further growth. Explains the risks of optimistic dispute resolution and how Chainlink's deterministic data streams and AI-hybrid oracles are reducing settlement times from hours to minutes.

Nic Carter examines the MicroStrategy Bitcoin sale market dispute and argues that Polymarket's entire market resolution system is structurally broken. He shows how outsourcing resolution to UMA creates a fundamental economic security failure: the UMA token market cap ($37M) is far smaller than the disputed market volume ($317M), making manipulation profitable and nearly undetectable. Carter compares Polymarket's approach to regulated venues like Kalshi and CME, which specify sources upfront, allow deferred settlement, and settle centrally without passing responsibility to anonymous token holders.

Evaluates whether multi-agent LLM architectures can resolve prediction market outcomes more accurately than single-model baselines. Tests independent aggregation and deliberative consensus against GPT-5 Nano, DeepSeek V3, and Llama-3.3-70B on 1,189 resolved questions from KalshiBench. Finds that confidence-weighted voting across agents edges past single models, while deliberation degrades accuracy — and proposes a hybrid system that auto-resolves unanimous high-confidence questions while flagging disagreements for human review.

An on-chain investigation revealing how a single entity extracted over $5m from Polymarket's crypto markets using 350+ accounts. The piece walks through the manipulation mechanism — accumulating positions on Polymarket then moving Binance's BTC price during settlement windows — and traces the fund flow through Hyperliquid wallets to Binance. Includes specific suggestions for TWAP-based settlement and account-level enforcement.

XO Labs publishes a framework for thinking about prediction market resolution as a spectrum rather than a single mechanism. The paper identifies the resolver trilemma between efficiency, decentralization, and security, using the Ukraine minerals market whale tax incident on Polymarket as a case study in token-weighted governance failure. They propose a layered architecture (XO Oracle) that routes each market to the level of security it actually requires.

Uses the $292M Kelp DAO rsETH bridge exploit to motivate why crypto-native parametric cover is needed, then compares HIP-4's binary event contract structure to CDS, catastrophe bonds, reinsurance sidecars, and weather derivatives. Argues that when outcome contracts share margin with underlying exposure on the same execution layer, HIP-4 unlocks a market two orders of magnitude larger than current DeFi insurance.

A formal comment to the CFTC's proposed rulemaking on prediction markets, submitted by an HKS researcher. Proposes a four-dimensional framework for classifying event contracts (information structure, manipulation economics, social utility, repugnance), reframes insider trading into three distinct patterns (outcome influence, duty breach, information advantage), and analyzes resolution integrity through three documented Polymarket/Kalshi case studies.

Catalog of 14 resolution failures over 18 months at Polymarket and Kalshi affecting over $500M in volume, including the $242M Zelenskyy-suit market, the $120M TikTok ban, and the $47M Cardi B halftime market that resolved yes on Polymarket and no on Kalshi. Groups the failures into four patterns: vague criteria, decentralized oracle capture (UMA token cap below disputed volume), centralized operator discretion, and cross-platform divergence. Useful reference for anyone designing resolution mechanisms or underwriting oracle risk.

Launch post for sdk.markets, a toolkit for creating parimutuel prediction markets on arbitrary questions, built on Base. Argues CLOB infrastructure does not fit thin community markets and that parimutuel pools are simpler and fairer when there is no natural counterparty. Details design choices that address parimutuel's classic 'wait and see' sniping problem (short answer windows, snapshot locking, DPM-style pricing) and three resolution modes: single admin, multi-admin consensus, and an AI oracle that resolves from arbitrary URLs.

Uses Kalshi's "mentions markets" (contracts that pay off if a specific word is said at a press conference) to illustrate a structural problem: prediction markets require crisp binary boundaries, but reality rarely provides them. Disputes over whether Cardi B "performed" at the Super Bowl, whether Zelenskiy "wore a suit," and what counts as a "word" show that platforms need linguists and philosophers as much as traders.

Walks through UMA's Optimistic Oracle pipeline as used by Polymarket: assertion, liveness period, dispute, and DVM token-holder voting. Uses the $240M Zelenskyy suit market as a case study where semantic ambiguity ('Is a militarized black outfit a suit?') triggered a disputed resolution ultimately decided by whale token holders, illustrating how the system's 'decentralized courtroom' handles edge cases.

Catalogs eight distinct arbitrage strategies available on prediction markets: classic YES+NO mispricing, cross-platform, range, conditional, time, hedged, resolution, and orderflow arbitrage. Each type includes concrete examples with dollar amounts and specific risk factors to watch for.

Argues the hardest PM problem isn't pricing but deciding what actually happened. Proposes cryptographically-committed LLMs as resolution judges—trading human bias and conflicts of interest for more tractable technical vulnerabilities. Cites Polymarket disputes (Venezuela election, Ukraine map, government shutdown) as evidence current systems fail at scale.

Examines a $10 million Polymarket dispute over whether events in Venezuela constitute an 'invasion,' exposing how semantic ambiguity in market resolution criteria can create massive financial consequences. Raises fundamental questions about who controls truth determination when contract language is open to interpretation.

Critiques Kalshi's 2025 measles cases market as an example of prediction markets being applied to inappropriate domains. Argues that turning a public health crisis into a speculative instrument is ethically questionable and reflects poorly on the industry's judgment about which events deserve tradeable contracts.

Comprehensive technical typology covering how prediction markets differ in outcome resolution (oracle-based vs self-resolving), trading mechanics, and design tradeoffs. Draws from established platforms and emerging ones to help users and developers navigate the ecosystem's diversity and associated risks.

Argues the Oracle Problem for prediction markets is actually a definition problem, not a trustlessness problem. Lists 12+ Polymarket controversies where contention arose from ambiguous event definitions, not oracle failures. Proposes that proprietary centralized oracles with platform stake are preferable to decentralized mercenary capital.