What is the Avellaneda-Stoikov strategy?

It is a mathematical market-making model that optimizes bid and ask prices based on inventory risk, volatility, and time horizon.

Is market making profitable for retail traders?

Sometimes, but it is competitive. Profitability depends on spreads, fees, inventory management, and market conditions.

What is inventory risk?

Inventory risk is the chance that you accumulate a position in an asset that then moves against you before you can offload it.

Can I lose money with Avellaneda market making?

Yes. Adverse selection, volatility spikes, and high fees can all cause losses. Start with paper trading.

Which pairs work best?

Pairs with consistent volume, tight natural spreads, and relatively stable prices tend to work better. Stablecoin pairs are common starting points.

Hummingbot Avellaneda: Crypto Market Making Basics

Risk Disclaimer: This content is for educational purposes only. Trading involves significant risk of loss. Past performance does not guarantee future results. Always do your own research before using any trading tool or strategy.

Market making sounds like free money. You place bids and asks, collect the spread, and repeat. In reality, it is one of the most competitive and risky forms of algorithmic trading. The Avellaneda-Stoikov strategy, available in Hummingbot, is a popular way for retail traders to experiment with market making. This guide explains how it works and how to use it responsibly.

What Is Market Making?

Market makers provide liquidity by continuously quoting buy and sell prices. They profit from the spread between bid and ask, but they also take on inventory risk. If prices move sharply, the market maker can be left holding a losing position.

The Avellaneda-Stoikov Model

The Avellaneda-Stoikov model is a classic market-making framework from quantitative finance. It calculates optimal bid and ask prices based on:

Current mid price: The average of best bid and best ask
Inventory position: How much of the asset you currently hold
Volatility: How much the price fluctuates
Time horizon: How long until the strategy rebalances
Risk aversion: How aggressively you want to reduce inventory risk

The strategy skews quotes to encourage inventory to return toward a target level. If you hold too much of the asset, it lowers the ask price to sell more. If you hold too little, it raises the bid price to buy more.

Setting Up Hummingbot

If you have not installed Hummingbot, start with Docker:

docker pull hummingbot/hummingbot:latest
docker run -it --name hummingbot hummingbot/hummingbot:latest

Inside Hummingbot, create the Avellaneda strategy:

create

Select avellaneda_market_making and follow the prompts.

Key Parameters

Parameter	What It Does	Typical Range
`market`	Trading pair such as BTC-USDT	Depends on exchange
`execution_timeframe_mode`	How often to refresh quotes	Infinite or from_date/to_date
`order_amount`	Size of each quote	Based on your capital
`risk_factor`	Controls inventory skew aggression	0.5 to 2.0
`order_refresh_time`	How often to cancel and replace orders	10 to 60 seconds
`minimum_spread`	Lowest spread you are willing to quote	Above trading fees
`volatility_buffer`	Safety margin for volatility estimates	0.1 to 0.5

A Conservative Starting Configuration

For a first experiment on a stablecoin pair, consider:

Order amount: small, such as $10 per side
Risk factor: 0.5 to 1.0
Order refresh time: 30 seconds
Minimum spread: 0.2% or higher
Paper trading enabled

This configuration will not make much money, but it will help you understand how the strategy behaves without large risk.

Paper Trading First

Hummingbot has a paper trading mode. Always enable it for your first runs:

paper_trade_mode_enabled: true

Watch how the bot reacts to price moves, how inventory shifts, and how fees accumulate. Run paper trading for at least a week before considering live capital.

Risks to Understand

Adverse Selection

Other traders may pick off your quotes when they have better information. You end up buying before a drop or selling before a rally.

Inventory Risk

If the market trends in one direction, your bot will accumulate the losing side of the trade. Without inventory controls, this can lead to large losses.

Fee Erosion

High trading frequency means high fees. Make sure your quoted spread comfortably covers exchange fees and expected slippage.

Volatility Spikes

Sudden volatility can move prices faster than the bot can adjust quotes. Kill switches and position limits are essential.

Monitoring

Track these metrics during paper and live trading:

Total P&L
Inventory ratio
Average spread captured
Fees paid
Number of trades
Maximum drawdown

When Market Making Makes Sense

Market making works best when:

Spreads are wide enough to cover fees
Volatility is moderate
You have enough capital to absorb inventory swings
You can monitor the bot closely

It works poorly when:

Spreads are razor-thin
Markets are trending strongly
Fees are high
You cannot react to problems quickly

Inventory Skew in Action

Imagine you run the Avellaneda strategy on BTC-USDT with a target inventory split of 50% BTC and 50% USDT. If a series of buys leaves you holding 70% BTC, the model adjusts quotes to encourage selling. Your ask price moves closer to the mid price, making it more likely to execute, while your bid price moves lower, discouraging further buying.

This dynamic rebalancing is the core of the Avellaneda-Stoikov model. The risk_factor controls how aggressively the bot skews. A higher risk factor means faster inventory correction but also narrower spreads and potentially lower profitability per trade.

Kill Switch and Circuit Breakers

Market making without a kill switch is dangerous. Configure Hummingbot to stop trading when:

Inventory skew exceeds a maximum threshold, such as 80% of capital in one asset.
Total P&L drawdown exceeds a predefined limit, such as 5% of capital.
Volatility spikes above a threshold measured by recent price range.
The exchange API returns repeated errors.

Test these kill switches in paper trading. In a live market, emotions and latency can prevent you from reacting manually.

When to Stop Market Making

Stop or reduce the strategy when:

Spreads tighten below your fee structure plus a safety margin.
The market enters a strong directional trend that creates one-sided inventory.
Exchange fees increase or rebates change.
You cannot monitor the bot for an extended period.

Market making is an active strategy, not a set-and-forget system. Constant monitoring and parameter tuning are required.

Bottom Line

The Avellaneda-Stoikov strategy is a powerful introduction to algorithmic market making. Hummingbot makes it accessible to retail traders. But accessibility is not the same as safety. Market making requires careful parameter tuning, inventory management, and constant monitoring.

Start with paper trading, use conservative settings, and never allocate capital you cannot afford to lose.