Building a Smarter Crypto Market Maker with Avellaneda–Stoikov

2 hours ago

Market making involves quoting both sides of the order book to profit from spreads, but faces inventory risk from accumulated positions.
The Avellaneda-Stoikov (AS) model dynamically adjusts quotes based on inventory, volatility, and risk tolerance to balance spread income and inventory risk.
The AS model calculates a reservation price that shifts from the mid-price depending on inventory and volatility, and an optimal spread influenced by risk aversion, volatility, and liquidity.
Implementation includes initialization of parameters, snapshot callback for order management, and finalization for cleaning up orders and logging.
Backtesting on BTC/USDT perpetual contracts uses DolphinDB's engine, with potential extensions like dynamic parameter estimation and multi-asset adaptation.

Hasty Briefsbeta