Why Prediction Markets Need Purpose-Built AMM Curves

A 50/50 market that shifts to 90/10 creates a massive liquidity trap. Generic AMMs like Uniswap v3 force LPs to post capital on the 'No' side that will likely never be used, crippling capital efficiency.

• ~80% of capital can be locked in a dead outcome.
• LPs face asymmetric, unbounded loss on the long tail.
• Creates perverse incentives to withdraw liquidity during volatility.

Purpose-built curves like the LMSR (used by Polymarket, Augur v2) price based on the cost of changing the probability distribution. Liquidity is a single, global pool that efficiently prices all outcomes simultaneously.

• Capital efficiency is outcome-agnostic; liquidity isn't tied to a specific price tick.
• Provides continuous, smooth pricing from 0% to 100% probability.
• The cost function guarantees bounded loss for market makers.

Concentrated liquidity AMMs require LPs to manually manage positions across hundreds of price ticks between 0 and 1. This creates fragmented, thin liquidity at critical probability points (e.g., 49% vs 51%), leading to high slippage on meaningful bets.

• Market resolution creates a winner-takes-all rush across the 0.5 threshold.
• Oracle latency of ~2-12 seconds allows for profitable front-running.
• Manual management overhead kills passive liquidity provision.

Systems like Uma's Optimistic Oracle or Chainlink Functions resolve markets off-chain, while the AMM curve automatically redistributes liquidity from losing outcomes to winners. This eliminates tick management and front-running vectors.

• Liquidity auto-compounds into the winning outcome, improving depth.
• Slippage is predictable and derived from the cost function, not fragmented ticks.
• Enables passive, hands-free liquidity provision for LPs.

In a prediction market, the asset expires worthless for 50% of holders. Generic AMMs treat this as a perpetual pair, forcing constant LP rebalancing as time decay accelerates. Settlement requires an oracle call and a chaotic, multi-step exit process for LPs.

• LP returns are dominated by volatile fee capture, not the fundamental yield of holding the winning asset.
• Post-resolution, LPs are left with a basket of worthless tokens, requiring manual unwinding.
• Creates systemic risk if the settlement oracle fails.

Separate the long-lived collateral (e.g., USDC) from the conditional outcome tokens. A purpose-built AMM trades these conditional claims. Upon resolution, the winning tokens redeem for collateral automatically; losers expire. This is the financial equivalent of a Merkle tree.

• Collateral is preserved and automatically routed to winners.
• LPs provide liquidity to the collateral pool, earning fees without time-decay risk.
• Settlement is atomic and trust-minimized via the smart contract.

In a binary market, LPs face maximum loss if they're wrong. A constant product curve forces them to take the opposite side of every trade, guaranteeing a 100% loss on the winning outcome. This is a structural disincentive that kills liquidity for niche events.

• Key Benefit 1: Purpose-built curves (e.g., logit, LMSR) cap LP loss at the provided liquidity.
• Key Benefit 2: Enables passive, diversified market-making akin to Polymarket's liquidity model, not directional betting.

A prediction market AMM must adjust its pricing curve based on time to resolution and trading volume. Early in an event, the curve should be shallow to attract information; near resolution, it must become steep to reflect certainty and prevent last-second arb.

• Key Benefit 1: Mimics traditional bookmaker models, optimizing for fee capture across the event lifecycle.
• Key Benefit 2: Automatically manages implied probability sensitivity, a feature absent in Uniswap or Curve.

A constant product curve requires ~4x the capital to achieve the same depth as a purpose-built curve for a binary outcome. This results in order-of-magnitude higher slippage for traders, making small markets unusable.

• Key Benefit 1: LMSR (Logarithmic Market Scoring Rule) provides constant liquidity depth regardless of price, a concept pioneered by Robin Hanson.
• Key Benefit 2: Enables high-resolution markets (e.g., "Will X win by 5-10%?") without exponential liquidity fragmentation.

The AMM contract must natively integrate a dispute resolution mechanism, not just a price feed. This requires a modular oracle design (e.g., UMA's Optimistic Oracle, Chainlink Functions) to finalize events and distribute pools.

• Key Benefit 1: Eliminates the need for a separate, trusted settlement layer, reducing protocol risk surface.
• Key Benefit 2: Creates a composable primitive for conditional tokens and derivatives, unlike monolithic designs.

A standard ERC-20 LP position from a generalized AMM cannot be natively interpreted as a claim on a specific outcome. This breaks DeFi composability for hedging, indexing, or using prediction shares as collateral.

• Key Benefit 1: Purpose-built AMMs mint non-fungible outcome tokens (like Gnosis Conditional Tokens), making them legible assets across DeFi.
• Key Benefit 2: Enables automated market makers for derivatives on top of the base layer, creating a full-stack prediction ecosystem.

Architects must expose curve parameters (e.g., liquidity sensitivity, fee decay rate) as governance levers. This allows a single AMM to service markets with different volatility profiles (e.g., sports vs. politics) without redeployment.

• Key Benefit 1: Protocol-owned liquidity can be optimally allocated across market categories based on risk-adjusted returns.
• Key Benefit 2: Creates a data moat; optimal parameters for event types become a learned, defensible advantage.