Prediction markets transform opinions into prices. When shares for "X will happen" trade at $0.70, the market collectively estimates a 70% probability. This makes prediction markets "economically-backed sources of truth" with real money at stake.
What Are Prediction Markets?
Prediction markets are platforms where users bet on the outcomes of future events using cryptocurrency. They operate through smart contracts that encode market rules, ensuring transparency and automatic settlement. Users buy and sell shares that pay out based on whether predictions prove correct.
The core innovation: prices reflect probability. If YES shares for "Bitcoin hits $100k by December" trade at $0.45, the market believes there's roughly a 45% chance of that outcome.
How Share Pricing Works
In binary markets (YES/NO outcomes):
- YES shares pay $1 if the event occurs, $0 if it doesn't
- NO shares pay $1 if the event doesn't occur, $0 if it does
- Share prices between $0.01 and $0.99 reflect implied probability
- YES + NO prices always equal approximately $1 (minus spread)
Pricing Mechanics: LMSR & Order Books
Different prediction markets use fundamentally different pricing engines. Understanding these mechanics matters because they determine liquidity depth, capital efficiency, and manipulation resistance.
Logarithmic Market Scoring Rule (LMSR)
Invented by Robin Hanson, LMSR is the automated market maker designed specifically for prediction markets. Instead of relying on order books, a single equation sets prices based on outstanding shares:
- Cost function: C(q) = b * ln(sum of e^(qi/b)) where b is a liquidity parameter and qi represents shares outstanding per outcome
- Key property: Bounded loss for the market maker — maximum subsidy is b * ln(n) where n is the number of outcomes
- Trade-off: Higher b = deeper liquidity but higher maximum loss; lower b = thinner markets but cheaper to operate
- Used by: Gnosis, Augur, and academic prediction markets
LMSR differs from Uniswap-style constant-product AMMs. LMSR guarantees bounded loss and is designed for events with known resolution dates. Constant-product AMMs are designed for continuous trading of fungible assets. Using a standard xy=k AMM for prediction markets creates poor pricing at probability extremes (near 0% or 100%).
CLOB (Central Limit Order Book)
Polymarket and Kalshi use traditional order books where makers post bids/asks and takers match against them. This mirrors equity exchanges and provides tight spreads when volume is high, but can leave thin markets for niche events.
Calibration & Forecasting Accuracy
The central claim of prediction markets is that prices reflect true probabilities. Empirical data now validates this:
| Metric | Finding | Source |
|---|---|---|
| Polymarket calibration | Events priced at 70% resolved YES approximately 70% of the time (>94% calibration accuracy across resolved markets) | 2024 US election cycle |
| Election volume | $3.7B cumulative volume on 2024 US presidential election markets alone | Polymarket |
| Polls vs markets | Prediction markets outperformed polling averages in 85%+ of state-level races in 2024 | Academic studies |
| Information speed | Markets update within minutes of news events; polls take days to reflect new information | Real-time analysis |
Polls measure stated preferences — what people say they'll do. Markets measure revealed preferences backed by real money. The $3.7B staked on the 2024 election meant participants had billions of dollars of incentive to be accurate, not just opinionated. This "skin in the game" is why well-calibrated prediction markets consistently outperform expert panels and polling aggregators.
Futarchy & Governance Markets
Beyond betting, prediction markets are being used for organizational decision-making — an idea called futarchy, proposed by Robin Hanson: "vote on values, bet on beliefs."
MetaDAO (Solana)
MetaDAO pioneered on-chain futarchy for protocol governance. Instead of traditional token voting:
- Conditional markets: For each proposal, two markets trade: "token price if proposal passes" vs "token price if proposal fails"
- Decision rule: If the "pass" market prices higher, the proposal is accepted — the market collectively believes it adds value
- Advantage: Eliminates voter apathy and plutocratic voting. Anyone with a view can express it with capital, and the market aggregates dispersed knowledge
- Track record: Multiple governance decisions executed via futarchy since 2024
Gnosis & Conditional Tokens
Gnosis developed the conditional token framework (ERC-1155 based) enabling prediction markets that branch on multiple conditions. This infrastructure powers governance-adjacent markets where outcomes depend on decisions yet to be made.
Market Types
Binary Markets
Simple YES/NO outcomes. Examples: "Will ETH be above $5,000 on January 1?" or "Will the Fed cut rates in March?" These are the most common and most liquid market types.
Categorical Markets
Multiple discrete outcomes. Example: "Which party wins the 2028 presidential election?" with shares for Democrat, Republican, Independent, and Other. Each outcome trades separately, with total probabilities summing to 100%.
Scalar Markets
Range-based outcomes. Example: "What will Bitcoin's price be on December 31?" with payouts calculated based on where the final price falls within a specified range.
Platform Models: Order Book vs AMM
| Aspect | Order Book (Polymarket) | AMM/Pool (Azuro) |
|---|---|---|
| Liquidity Source | Active traders placing orders | Pooled deposits from LPs |
| Price Discovery | Bid/ask spread determines price | Algorithm based on pool ratios |
| Best For | High-volume, high-interest events | Long-tail markets, sports betting |
| LP Role | Market makers set prices | Passive deposits earn fees |
| Risk Profile | Traders bear directional risk | LPs act as counterparty to bettors |
Major Platforms
Polymarket
The dominant crypto-native prediction market, Polymarket runs on Polygon and uses an order book model. Key characteristics:
- Focus: Politics, news events, crypto, culture
- Volume: Weekly volumes reached $3.8B in late 2025
- Valuation: $8-9B after $2B investment from ICE (NYSE parent)
- Oracle: UMA's Optimistic Oracle for dispute resolution
- Upcoming: POLY token launch expected with US market re-entry
Polymarket is currently unavailable to US users due to a 2022 CFTC settlement. The platform operates from outside the US but plans to re-enter the market with regulatory compliance.
Kalshi
The regulated alternative, Kalshi operates as a CFTC-regulated exchange available to US users:
- Focus: Primarily sports (90%+ of volume), plus politics and economics
- Valuation: $11B after raising $1B in funding
- Market Share: ~47% of prediction market open interest
- Regulatory Status: First legally approved prediction market in the US
Azuro
A B2B protocol enabling anyone to build prediction market frontends:
- Model: Peer-to-pool where LPs provide passive liquidity
- Focus: Sports betting (58% football, 12% basketball, 11% tennis)
- LP Returns: Historically ~20% APY, currently ~12%
- Token: $AZUR governance token
- Innovation: Third-party apps can launch without bootstrapping liquidity
Market Growth & Metrics
Weekly Volume: $3.8B (660% increase from January)
Weekly Transactions: 12.7M (from 1.2M)
Open Interest: $700M+ (below $900M election peak)
Monthly Volume: Approaching $10B in Q4 2025
| Platform | Open Interest Share | Volume Share | Primary Category |
|---|---|---|---|
| Polymarket | 44% | ~35% | Politics (26%), Sports (37%), Crypto (25%) |
| Kalshi | 47% | ~40% | Sports (90%+) |
| Opinion | ~5% | ~43%* | Various (*incentive-driven) |
| Azuro | ~4% | ~5% | Sports (95%+) |
Oracle & Resolution Mechanisms
Prediction markets need reliable ways to determine outcomes. Different approaches include:
Optimistic Oracles (UMA)
Polymarket uses UMA's system where outcomes are proposed and accepted unless disputed within a challenge period. Disputers must stake tokens, creating economic incentives for honest reporting.
Decentralized Resolution
Some platforms use token-holder voting or multi-party consensus to resolve disputed outcomes, though this can introduce delays and governance attack risks.
Centralized Resolution
Kalshi, as a regulated exchange, uses internal resolution with regulatory oversight—faster but requiring trust in the operator.
Oracle manipulation remains a key risk. If the resolution mechanism can be compromised, markets can be settled incorrectly. This is why Polymarket markets occasionally face controversial resolutions that require human judgment beyond simple data feeds.
Use Cases Beyond Betting
Information Aggregation
Prediction markets aggregate dispersed information into prices. During elections, market prices often prove more accurate than polls because traders have financial incentives to be right, not just to voice opinions.
Hedging Instruments
Businesses can hedge against regulatory, political, or macroeconomic outcomes. A crypto company worried about adverse regulation could buy shares in "Strict crypto regulation passes" to offset potential losses.
Media Integration
News outlets increasingly embed prediction market odds alongside traditional polling data. Google Finance, Robinhood, and major media organizations are integrating market data as sentiment indicators.
AI Agent Markets
Emerging use case: AI agents autonomously creating markets, providing liquidity, and trading based on information analysis—potentially making prediction markets more efficient and comprehensive.
Risks & Challenges
Liquidity Constraints
Thin markets enable manipulation and create wide spreads. A well-funded actor can move prices significantly in illiquid markets, making them less reliable as information sources.
Regulatory Uncertainty
Prediction markets occupy an unclear regulatory space—part gambling, part derivatives, part information services. Different jurisdictions treat them differently, creating operational complexity.
User Experience
Wallet management, gas fees, and crypto complexity still deter mainstream users. Kalshi's regulated status enables traditional payment methods, giving it an advantage for US retail adoption.
The Parlay Problem
Sports betting especially faces challenges with leveraged "parlay" bets that create outsized liabilities for liquidity providers, making sustainable LP returns difficult to achieve.
Investment Implications
What to Watch
- Token launches: Polymarket's POLY token could be a major 2026 catalyst
- Sports betting integration: 10% CAGR projected through 2030 ($198B market)
- Regulatory developments: US market access is the key unlock
- LP economics: Can platforms sustain attractive yields for liquidity providers?
Evaluation Framework
- Volume concentration: Heavy reliance on single events (elections) creates volatility risk
- User retention: Polymarket users are often one-time speculators; Azuro sees more recurring engagement
- Oracle reliability: Resolution mechanism quality directly impacts platform trust
- Regulatory positioning: Compliance infrastructure determines geographic expansion potential
Prediction markets are evolving from niche crypto applications to mainstream financial infrastructure. The combination of regulatory progress (Kalshi), institutional investment (ICE/Polymarket), and growing media adoption suggests the sector is at an inflection point. Key risks remain around liquidity, regulation, and oracle reliability.