Pantera Research Lab's empirical study of the $7.8B short-term crypto prediction market on Polymarket and Kalshi, analyzing fee economics, maker rebate structures, and settlement-driven trading patterns. Finds Polymarket's 5-minute market is 86% bot-driven by wallet count, with Binance spot volume spiking 12-17x in the final seconds of still-even cycles, consistent with settlement-driven strategies.

Examines how a $5.7 million Polymarket bet on whether a gamer said "donk" sparked a revolt over the platform's optimistic oracle resolution system. When one firm's token-weighted vote effectively decided the outcome, traders accused Polymarket of becoming a popularity contest with a wallet rather than a truth machine — exposing deep tensions in how decentralized prediction markets handle ambiguous outcomes.

Polymarket matches orders off-chain for speed but settles on-chain for finality. This paper identifies a consistency gap they call Ghost Fills — orders matched off-chain that fail during on-chain settlement. Across 1.95M reverted transactions, the authors document four attack vectors that let traders selectively invalidate orders, realizing $1.49M in profit with 24.3% of filled orders reverting during peak hours.

Investigates how easy it is to rig prediction markets on Rotten Tomatoes movie scores — a handful of compromised critics could swing scores enough to cash out on Kalshi. Chronicles multiple real insider trading cases: the Khai Rated critic who tanked Alien: Earth scores, the Maduro raid trader, MrBeast's video editor, the Google employee arrested for betting on search trends. Examines why CFTC enforcement has weakened under Trump while platforms like Kalshi struggle to self-police.

An accessible explainer from Scott Duke Kominers on how prediction markets aggregate information through market prices. The piece covers the core mechanism, benefits over polls, and key design challenges around participation, insider trading, and manipulation.

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.

Seva Gunitsky argues that prediction markets create three distinct dangers for global conflict: military insider trading at scale, manipulation of outcome resolution, and propaganda advantages for autocratic regimes. The article walks through real cases including the Israeli Air Force betting scandal and the 'hair dryer problem' — cases that challenge the idealistic framing of prediction markets as truth machines.

Robin Hanson responds to a CFTC call for comments by arguing prediction markets deserve the same regulatory treatment as other information institutions like journalism and academia. Drawing parallels between six common harms shared across all information systems, from insider trading to manipulation, he contends markets should be approved by default and restricted only on clear evidence of specific harm. The piece makes the economist's case that the information value of prediction markets justifies a lighter regulatory touch than traditional gambling law.

Defense of prediction markets that reframes the moral critique as a critique of capitalism itself. Litvin walks through the standard objections (gambling, insider trading, manipulation, slot-machine durations) and pairs each with a larger-scale analog in traditional finance: the $950M oil ceasefire trades on CME, LIBOR, accredited investor rules, dollar debasement. Argues that the legal line between gambling and investing collapses under scrutiny and that prediction markets are simply a more legible version of dynamics already accepted everywhere else.

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.

Argues play-money prediction markets are accurate only because they operate in a low-manipulation regime, and this accuracy is self-undermining: the more important they become, the more valuable they are to manipulate. Contrasts the manipulation resistance of real-money markets against the economic noise they introduce, including opportunity cost, hedging distortion, and participation friction.

Traces the history of prediction markets from 1419 Vatican papal elections through the Iowa Electronic Markets to the Polymarket era, arguing the sector is at an inflection point. Surveys the insider trading scandals (Musk tweets, French elections), moral hazard concerns (assassination markets), and the wave of new entrants (Robinhood, DraftKings, Crypto.com, FanDuel) that signal mainstream adoption. Concludes that prediction markets are evolving, not decaying, but need regulatory clarity and structural reform to mature.

Documents a pattern of insider trading on prediction markets, from wallets that profited $1.2 million on the timing of US strikes on Iran to trades linked to classified intelligence. Compares how Kalshi's KYC-based surveillance and Polymarket's pseudonymous blockchain create different enforcement challenges. Argues platforms should reconsider contract offerings before regulators act.

Written during the US-Israel strikes on Iran, examines whether prediction markets on armed conflicts are net informational goods or perverse incentive engines. Dissects the IDF insider trading case where soldiers traded Polymarket positions before strikes, the CFTC's regulatory stance, and the divergent approaches of Kalshi (regulated, avoids conflict markets) versus Polymarket (offshore, lists them freely). Argues the information value is real but the moral hazard is structurally underpriced, and proposes guardrails including delayed settlement and conflict-of-interest screens.

Argues that prediction markets are financial instruments, not gambling, by examining Polymarket's architecture across multiple layers: peer-to-peer order book mechanics, information aggregation through skin-in-the-game pricing, hedging use cases, and UX design that suppresses gambling patterns. Contrasts the exchange model with the house-edge casino model to argue the gambling label stems from outdated legal frameworks.

Argues that leverage solves, rather than amplifies, prediction market problems. In a 1x market, only whales can move prices because the barrier is capital, not insight; at 10x, thousands of smaller traders can collectively contest a mispricing. Addresses gap risk (binary resolution makes platform-native margin dangerous) by describing a temporal arbitrage approach where leveraged markets close before event resolution. Also proposes a vault-based yield layer where LPs earn returns from trading activity rather than directional outcome exposure.

Argues that nearly any prediction market tied to a public figure's actions, tenure, or appearances implicitly embeds assassination as a resolution path — what the authors call 'kinetic intervention.' Uses the Charlie Kirk assassination and subsequent Kalshi market voiding as the central case study. Warns that blanket void-on-death rules can backfire by incentivizing violence from losing bettors, and proposes that platforms hire geopolitical risk officers to evaluate resolution wording, monitor anomalous betting signals (BETINT), and build early-warning capacity before tragedy occurs.

An academic paper from Stanford and Singapore Management University studying how prediction market participants can manipulate the settlement process itself — the mechanism by which outcome data feeds into market resolution. Extends the manipulation literature beyond the trading phase to examine oracle-level attacks, where traders profit by influencing or anticipating biased settlement rather than by trading on superior information. Formalizes the conditions under which settlement manipulation is profitable and discusses implications for oracle design, dispute resolution frameworks, and market integrity.

Examines Clinton and Huang's research on 2024 election market accuracy, finding PredictIt at 93%, Kalshi at 78%, and Polymarket at 67%, while also documenting significant cross-platform price divergences for identical contracts near Election Day. Raises concerns about Kalshi's media partnerships with CNN and CNBC, arguing they create incentives for sensational coverage of market movements and potential manipulation of thin markets.

Examines how to distinguish wash trading from legitimate market making on Polymarket using network analysis. Wash traders exhibit homophily, trading only within their collusive group, while market makers trade indiscriminately with diverse counterparties. Describes an algorithm developed by Columbia researchers that identified a cluster of 200 wallets generating $113 million in volume with just $57.86 in aggregate losses.

Large-scale field experiment testing prediction market manipulation across 817 markets. Randomly shocked prices and tracked effects over 60 days with hourly data. Finds markets can be manipulated with effects persisting for months, though they gradually fade. Markets with more traders, higher volume, and external probability estimates prove more resistant.