Order Books vs AMMs

The price of a prediction market contract is dictated entirely by how buyers and sellers are matched. In classical finance, this is done via a Central Limit Order Book (CLOB). In decentralized finance (DeFi), this is often handled by an Automated Market Maker (AMM).

Understanding the structural differences between a CLOB and an AMM will dictate which platforms you use and how you execute your trades.

1. Central Limit Order Book (CLOB)

A CLOB is the traditional method used by stock exchanges (NYSE, Nasdaq) and regulated prediction markets like Kalshi and the Polymarket CLOB API.

How a CLOB Works

A CLOB is simply a ledger that records all outstanding buy orders (Bids) and sell orders (Asks).

• Bids: The highest price someone is willing to pay to buy a share.
• Asks: The lowest price someone is willing to accept to sell a share.

When a new order arrives, the matching engine attempts to pair it with an existing order on the opposite side of the book. If the new order is a Market Order, it executes immediately against the best available price. If it is a Limit Order, it sits on the book until another trader is willing to match that specific price.

Pros & Cons of a CLOB

• Pros:
  • Zero slippage on Limit Orders (your order executes at your exact price, or not at all).
  • Highly capital efficient for active market makers.
  • Supports complex routing and institutional algorithms.
• Cons:
  • Requires active participants to provide liquidity (Market Makers). If the market makers leave, the spread widens, and trading grinds to a halt.
  • Vulnerable to "flash crashes" if the order book is thin.

2. Automated Market Makers (AMM)

An AMM replaces the traditional matching engine and human market makers with a mathematical formula running on a smart contract. Rather than trading against another human, you trade directly against the algorithm's pool of liquidity.

How an AMM Works

In an AMM, liquidity providers (LPs) deposit collateral into a smart contract pool. When a trader wants to buy a "Yes" share, the AMM mints the share, takes the trader's cash, and uses a deterministic mathematical function to adjust the price upward for the next buyer.

Because the AMM is an algorithm, it will always quote you a price, regardless of whether any other human wants to take the opposite side of your trade. The price is dictated purely by the ratio of "Yes" to "No" shares currently held in the pool.

The LMSR (Logarithmic Market Scoring Rule)

The original AMM algorithm designed specifically for prediction markets was the LMSR, popularized by Robin Hanson. It was designed to ensure that the market price precisely reflected the crowd's probability estimation, while bounding the maximum potential loss for the market creator.

While highly accurate for information discovery, the LMSR is computationally heavy for blockchain environments due to exponential functions, and requires the market creator to subsidize the initial liquidity pool (which they will structurally lose).

The CPMM (Constant Product Market Maker)

To solve the limitations of LMSR on the blockchain, DeFi protocols like Uniswap introduced the Constant Product Market Maker (CPMM). This formula is stunningly simple:

$$ x \times y = k $$

• x: The amount of "Yes" shares in the pool.
• y: The amount of "No" shares in the pool.
• k: A constant value that never changes during a trade.

If you buy a large amount of "Yes" shares, the pool's remaining supply of "Yes" decreases ($x$ goes down). To keep $k$ constant, the price of "Yes" rises along the curve. That is slippage.

The CPMM allows anyone to become a Liquidity Provider and earn trading fees, making the market self-sustaining without requiring a central entity to subsidize it.

Pros & Cons of an AMM

• Pros:
  • Always provides liquidity; you can trade 24/7 without waiting for a counterparty.
  • Democratizes market making (anyone can provide liquidity).
  • Highly resistant to censorship and manipulation (since the math is public).
• Cons:
  • Impermanent Loss: Liquidity providers can lose money if the market price trends strongly in one direction and resolves.
  • High Slippage: Large trades usually incur more slippage as you move along the CPMM curve.
  • Capital inefficient compared to a CLOB.

Which is Better?

The future of prediction markets is likely a hybrid approach.

Platforms like Polymarket initially launched purely on an AMM architecture but have increasingly shifted volume and liquidity incentives toward their high-performance CLOB. Today, institutional liquidity providers use the Polymarket CLOB via API, while retail users interact with an aggregated view of both the CLOB and the underlying AMM backup pool.

If you are executing institutional-sized trades or building automated bots, you will almost certainly use the resting limit orders of a CLOB to eliminate slippage.

If you are trading highly illiquid, niche markets, an AMM may be the only source of continuous quoting available.