Two fundamentally different systems power today's largest prediction markets. Polymarket runs on an Automated Market Maker (AMM). Kalshi runs a Central Limit Order Book (CLOB). The difference shapes everything: liquidity, spreads, price discovery speed, and who can profitably trade.

How an AMM works

An AMM uses a mathematical formula to set prices automatically. In Polymarket's case, the CLOB model (they actually migrated from pure AMM in 2023) uses a constant product formula adapted for binary outcomes. The key property: you can always trade, but the price moves against you as your order size grows. Large orders move the market more.

This creates a predictable liquidity profile. Liquidity providers deposit both YES and NO shares into a pool and earn fees from trading activity. They don't need to be active — the contract prices them automatically.

How an order book works

Kalshi's CLOB is more similar to a stock exchange. Buyers post bids ("I'll pay $0.40 for YES"), sellers post asks ("I'll sell YES for $0.42"), and the exchange matches them when prices cross. There's no automatic price — the market price is simply the last trade price.

Order books require active market makers to maintain liquidity. During high-volatility events (like a surprise jobs number), order books can experience "liquidity gaps" where spreads widen dramatically. AMMs don't gap — they just move the price along the curve.

What this means for you

For small trades: AMMs are usually better. The spread is predictable and you execute immediately. For large trades (>$50k notional): order books are usually better if there's deep liquidity, because you won't move the price curve. For fast-moving events: order books update faster because they rely on human market makers who can reprice in milliseconds based on news, whereas AMM prices change only when someone trades against the pool.