Three models of prediction markets with the same outcome but completely different architecture
Traditional betting: the bookie is everything
sets the odds, holds the liquidity, takes the other side of your bet. one entity controls the entire stack and if they don't like you winning, then they cut you off.
Prediction markets v1: remove the bookie
bettors match against each other through a CLOB or AMM (peer-to-peer). but now you have a different problem where you need someone on the other side of every position
this often leads to thin markets, wide spreads and illiquid tails
Azuro unbundles the bookie into three separate roles:
β data providers compete to set the most accurate odds and earn revenue for it
β LPs provide liquidity through the liquidity tree and earn yield from volume
β the vAMM connects both sides, settles everything onchain
no single entity controls odds, liquidity, and the market at the same time
this is why azuro can power many apps from one liquidity layer
the architecture was designed to separate concerns from the start and not bolt decentralization onto a bookie model after the fact.