Avellaneda–Stoikov Quote Bands for Threshold Trading Signals

Summary

This document adapts the Avellaneda–Stoikov market making model into threshold based signals. It estimates a mid-price from the current and prior prices, then constructs bid and ask levels by offsetting that midpoint with a term based on gamma, volatility, time, and k. A further threshold determines when price is sufficiently below the bid level to signal a buy or sufficiently above the ask level to signal a sell. The supplied parameters include these model inputs and the threshold.

The text highlights parameter selection, liquidity, and execution latency as important limitations, and notes that transaction costs matter. However, although it describes a fee adjustment, the provided code sets the fee to zero. The code displays signals and enters long or short positions, but does not model inventory-aware quote placement or document realized market making performance. A BTC/USDT futures backtest configuration is included without reported results, so the document offers a signal recipe rather than evidence of profitability.

Key ideas

• The model uses the average of the current and previous prices as its midpoint estimate.
• Bid and ask levels are formed by subtracting or adding a parameterized spread term and a fee adjustment.
• A threshold beyond these levels triggers long or short entries.
• Results depend on parameter choices, liquidity, transaction costs, and execution speed.
• The source sets fees to zero and reports no backtest performance.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.