Market Making Strategy: How Market Makers Profit from Spreads 2026

What Is a Market Making Strategy?

A market making strategy is a trading approach where a firm continuously quotes both a buy price (bid) and a sell price (ask) for a financial instrument, profiting from the spread between them while managing the risk of holding inventory. Market makers don't bet on whether prices will go up or down - they earn from the act of providing liquidity.

Here's the basic mechanics. A market maker posts a bid of £100.00 and an ask of £100.02 on a stock. If one trader sells to the market maker at £100.00 and another buys from the market maker at £100.02, the market maker has earned £0.02 per share without taking any directional view. That £0.02 is the spread - the fundamental unit of market making revenue.

In practice, it's never that simple. The market maker rarely gets perfectly balanced flow where buys and sells arrive equally. Instead, they accumulate inventory - a position they didn't necessarily want - and must manage the risk that the price moves against them before they can flatten out. A market maker who buys 50,000 shares at £100.00 and then watches the price fall to £99.80 has just lost £10,000, which wipes out the spread earned on thousands of previous round trips.

This is why market making is fundamentally a risk management business, not a prediction business. The strategy works when spread income consistently exceeds the losses from adverse price movements and the costs of operating the technology required to quote competitively. In 2026, all major market making is electronic. Firms run automated systems that quote thousands of instruments simultaneously, updating prices multiple times per second in response to changing market conditions. For a deeper look at the mechanics of how exchanges process these quotes, see our guide to market microstructure.

---

How Market Makers Make Money

Market makers make money through three primary channels: spread capture, exchange rebates, and information advantage from observing order flow. The relative importance of each varies by asset class and market structure.

Spread capture is the core revenue source. Every time a market maker buys at the bid and sells at the ask, they capture the spread. On a liquid stock with a one-penny spread, this might be £0.01 per share. Multiply that across millions of shares traded per day and the numbers become significant. The key is achieving balanced flow - roughly equal buying and selling - so that inventory doesn't build up in one direction. When flow is balanced, spread capture is almost mechanical profit.

Exchange rebates add a second revenue stream. Most electronic exchanges operate on a maker-taker fee model. Liquidity providers (makers) who post resting limit orders receive a small rebate - typically a fraction of a penny per share - when their orders are filled. Liquidity takers who send marketable orders pay a fee. For a high-volume market maker, rebate income can amount to millions of pounds per year. Some market making strategies are specifically designed to maximise rebate capture, particularly on venues where the rebate is generous relative to the spread.

Information advantage from order flow is the subtlest source of edge. A market maker who quotes across thousands of instruments sees an enormous amount of order flow data. Patterns in that flow - which instruments are being bought or sold, in what size, and by what type of counterparty - contain information about short-term price direction. Large electronic market makers are sometimes described publicly as seeing a large share of aggregate equity flow; shares move over time. This is not the same thing as misusing client-order knowledge, which is addressed separately under market-manipulation and conduct rules; MM quoting decisions still react to statistical patterns in aggregated activity.

The costs that erode market making profits include adverse selection losses (trading against informed counterparties who know something the market maker doesn't), inventory risk (holding positions during unfavourable price moves), technology costs (co-location, hardware, engineering talent), and regulatory costs (compliance, reporting, capital requirements).

---

The Avellaneda-Stoikov Model

The Avellaneda-Stoikov model (2008) is the foundational academic framework for optimal market making. It answers the question: given uncertainty about future price movements, where should a market maker place their bid and ask quotes to maximise expected profit while controlling risk?

The model starts with a market maker who wants to quote on a single asset. The asset's mid-price follows a random walk (Brownian motion), and the market maker faces a stream of incoming buy and sell orders that arrive randomly. The market maker must decide, at each moment, how far from the mid-price to set their bid and ask quotes.

The core insight is that optimal quote placement depends on three factors:

Inventory. If the market maker has accumulated a long position (too many shares), they should lower both their bid and ask to encourage selling and discourage further buying. If they're short, they do the opposite. The quotes skew in the direction that pulls inventory back towards zero. The more extreme the inventory, the more aggressive the skewing.

Volatility. Higher volatility means greater risk from holding inventory, so the market maker should quote wider spreads. When the market is calm, spreads can tighten because the risk of a large adverse move is smaller.

Risk aversion. A more risk-averse market maker will quote wider spreads and skew more aggressively against their inventory. The model uses a parameter (gamma) that captures how much the market maker penalises variance in their P&L.

The model produces closed-form solutions for optimal bid and ask prices as functions of current inventory, time, volatility, and risk aversion. The reservation price - the market maker's indifference price given their current inventory - shifts away from the mid-price as inventory increases. The optimal spread around this reservation price widens as volatility or risk aversion increases.

In practice, no production market making system runs pure Avellaneda-Stoikov. The model makes simplifying assumptions - constant volatility, Poisson order arrivals, a single instrument - that don't hold in real markets. But the intuitions transfer directly. Every modern market making algorithm incorporates inventory-dependent quote skewing, volatility-adjusted spread widths, and some form of risk penalty. The model provides the intellectual scaffolding; firms build proprietary extensions that handle the real-world complexities of multiple instruments, correlated assets, discrete tick sizes, and non-constant volatility.

For those building market making systems, Avellaneda-Stoikov is the starting point, not the destination. Extensions by Guéant, Lehalle, and Fernandez-Tapia (2013) handle multi-asset market making. Cartea, Jaimungal, and Penalva provide a comprehensive treatment in their textbook on algorithmic trading and quantitative strategies.

---

Key Components of a Market Making Strategy

Every market making strategy, regardless of asset class, must solve four problems simultaneously: estimating fair value, calculating the spread, managing inventory, and handling orders. Get any one of these wrong and the strategy fails.

Fair Value Estimation

The market maker needs a real-time estimate of where the "true" price is. This is harder than it sounds. The mid-price (average of best bid and best ask) is the simplest estimate, but it's noisy and can be manipulated. More sophisticated approaches incorporate:

• Weighted mid-prices that account for the size at each level of the order book
• Microprice models that weight the bid and ask by their respective queue sizes
• Signals from correlated instruments (if futures move before the stock, the futures price is informative)
• Short-term predictive models that incorporate recent order flow, trade imbalance, and volatility

The quality of fair value estimation directly determines profitability. A market maker whose fair value estimate is consistently 0.1 pennies more accurate than competitors captures that edge on every trade.

Spread Calculation

The quoted spread around fair value determines the trade-off between volume and profit per trade. A tighter spread attracts more volume but earns less per fill and provides less cushion against adverse moves. A wider spread earns more per fill but may result in fewer executions.

Optimal spread width depends on volatility (wider when volatile), order arrival rates (tighter when flow is heavy and likely to be balanced), inventory (wider when inventory is elevated to reduce further accumulation), tick size (the minimum possible spread is one tick), and competition (no point quoting wider than the next-best market maker).

Inventory Management

This is the single most important risk management component. The market maker must keep inventory within acceptable bounds and actively work to return it towards a target level - typically zero or near-zero.

Inventory management techniques include quote skewing (adjusting bid and ask asymmetrically to attract flow that reduces inventory), hedging in correlated instruments (offsetting equity inventory with futures, for example), position limits and circuit breakers (hard limits that pull quotes entirely if inventory exceeds thresholds), and time-based urgency (becoming more aggressive about reducing inventory as the trading day progresses).

Order Management

The mechanics of how orders are submitted, modified, and cancelled matter enormously at scale. A market maker quoting on 5,000 instruments might generate millions of order messages per day. The order management system must handle quote updates in microseconds, manage queue position (cancelling and re-entering loses your place in the queue), handle partial fills correctly, comply with exchange-specific order type rules, and avoid self-trading (accidentally trading against your own resting orders on the same venue).

---

Inventory Risk Management

Inventory risk is the central danger in market making. Every share or contract a market maker holds represents exposure to an adverse price move, and the longer they hold it, the more likely that move becomes.

Consider a market maker on a mid-cap equity who accumulates a long position of 100,000 shares over the course of a morning. If the stock drops 0.5% before they can unwind, that's a loss of £50,000 on a £10 million position - far more than the spread income earned while building that inventory. One bad inventory blowout can wipe out weeks of spread profits.

Why inventory accumulates. In a perfectly balanced market, buy and sell orders arrive equally and the market maker's inventory stays near zero. But real markets are rarely balanced. A large institutional seller might lean on the market maker's bid for hours, creating a growing long position. Or a momentum move might trigger a cascade of buying that leaves the market maker deeply short.

Mean-reverting inventory targets. The standard approach is to target zero inventory and apply increasingly aggressive measures to return there. When inventory is small, gentle quote skewing is sufficient - nudging the bid and ask a fraction of a tick to encourage offsetting flow. When inventory is moderate, the market maker might hedge in a correlated instrument (buying index futures to offset a short equity position, for instance). When inventory is large and approaching limits, the strategy may cross the spread - sending marketable orders that guarantee execution but cost the spread - to reduce exposure immediately.

Hedging across instruments. Market makers rarely manage a single instrument in isolation. A firm making markets on 500 equities can offset inventory in individual names by trading sector ETFs, index futures, or baskets of correlated stocks. This cross-instrument hedging reduces net exposure without requiring that every individual position be flat. It also introduces basis risk - the risk that the hedge and the underlying don't move in perfect lockstep.

End-of-day flattening. Most market making firms aim to be flat or near-flat at the end of each trading day. Carrying overnight inventory exposes the firm to gap risk - the possibility that news breaks outside trading hours and the stock opens at a dramatically different price. As the close approaches, inventory management becomes more aggressive, and firms will pay more to flatten out.

---

Adverse Selection

Adverse selection is the market maker's most dangerous enemy. It occurs when the market maker trades against a counterparty who has better information about where the price is heading - and the market maker ends up on the wrong side.

Here's how it works. A market maker is quoting £100.00 bid and £100.02 ask on a stock. A large, sophisticated hedge fund has just completed proprietary research suggesting the stock is worth £102. The fund sends a buy order that hits the market maker's ask at £100.02. The market maker has just sold shares at £100.02 to someone who knows they're worth £102. As the information becomes public and the price rises, the market maker is sitting on a short position that's deeply underwater.

This is toxic flow - order flow from informed participants that systematically loses money for the market maker. The challenge is distinguishing toxic flow from benign flow in real time.

Detecting informed flow. Market makers use statistical methods to classify order flow. Signals that suggest a counterparty might be informed include aggressive order sizing (much larger than typical retail orders), clustering of orders ahead of news events, orders that arrive in patterns matching known algorithmic execution strategies used by hedge funds, and high fill rates on one side of the book (when someone is consistently lifting the ask, it often signals information).

Protecting against adverse selection. The primary defence is widening the spread for flow that looks informed and tightening it for flow that looks uninformed. This is what designated market maker programmes on some exchanges allow - the ability to provide different prices to different counterparty types. On public exchanges where all orders are anonymous, market makers widen spreads generally during periods when informed trading is likely (before earnings announcements, during unusual options activity, when correlated instruments are moving).

The adverse selection cost model. Academic research, particularly the Glosten-Milgrom model (1985), shows that the bid-ask spread must compensate for adverse selection even in a perfectly competitive market. The spread has an adverse selection component - the expected loss from trading against informed counterparties - and the market maker can't profitably quote tighter than this floor. Understanding and measuring the adverse selection component of the spread is a core research problem at market making firms. For more on these models, see our market microstructure guide.

---

Market Making Across Asset Classes

The core principles of market making - provide liquidity, earn the spread, manage inventory - apply universally, but the details differ significantly across asset classes. A market making strategy that works for equities won't transfer directly to options, crypto, or fixed income without substantial modification.

Equities

Equity market making is the most competitive and most technology-intensive segment. Spreads on liquid stocks are often at the minimum tick size (one penny in the US, variable in Europe), which means the profit per trade is tiny and volume is everything. Success depends on speed (microsecond quote updates), fair value estimation (incorporating signals from futures, ETFs, and correlated stocks), and scale (quoting on thousands of names simultaneously). The dominant firms - Citadel Securities, Virtu Financial, and a handful of others - have invested billions in infrastructure.

Options

Options market making adds the complexity of non-linear payoffs. An options market maker must manage delta, gamma, theta, and vega across thousands of contracts - different strikes, expiries, and underlyings. The volatility surface replaces the simple mid-price as the reference for fair value. Hedging is more complex because delta changes as the underlying moves (gamma risk), and implied volatility itself can shift. The computational demands are higher: repricing thousands of options contracts multiple times per second when the underlying moves requires serious computing power.

Crypto

Crypto market making has matured significantly since the early days of fragmented, unreliable exchanges. In 2026, firms like Jump Crypto, Wintermute, and DWF Labs make markets on centralised exchanges (Binance, Coinbase, Kraken) and on-chain through automated market making (AMM) protocols on DeFi platforms. AMMs like Uniswap use a fundamentally different mechanism - a mathematical formula (the constant product formula, x * y = k) replaces the traditional order book. Liquidity providers deposit tokens into pools and earn fees proportional to their share of the pool. The risks are different too: impermanent loss replaces traditional inventory risk, and smart contract risk adds a dimension that doesn't exist in traditional markets.

Fixed Income

Fixed income market making - bonds, rates, credit - is the least automated of the major asset classes, though this is changing rapidly. Bonds don't trade on central limit order books the way equities do. Instead, most bond trading is request-for-quote (RFQ), where a client asks multiple dealers for a price on a specific bond. The market maker must price the bond quickly (often using models that reference benchmark curves and credit spreads), decide whether to commit capital, and manage the resulting inventory. The challenge is that individual bonds are far less liquid than stocks - a single corporate bond might trade only a few times per week.

---

Technology Requirements

Market making in 2026 is as much a technology business as a trading business. The infrastructure requirements scale with the competitiveness of the market you're making.

Speed and latency. In equity and futures market making, the ability to update quotes in single-digit microseconds is table stakes. Firms use co-located servers positioned inside exchange data centres, direct market data feeds (bypassing slower consolidated feeds), custom networking hardware (kernel-bypass networking, FPGA-based network interface cards), and trading logic implemented on FPGAs or in highly optimised C++ with careful memory management and zero-allocation paths. For a detailed look at the technology arms race, see our high frequency trading guide.

Risk systems. A market maker's risk engine must calculate position exposure, P&L, and inventory metrics in real time across every instrument being quoted. If the risk engine is slow, the firm is quoting blind - offering prices without knowing their true exposure. Production risk systems run on pre-allocated memory with lock-free data structures, updating positions and Greeks (for options) in microseconds.

Pricing infrastructure. Fair value models must run fast enough to keep up with market data. For equities, this might be a lightweight regression model. For options, it's a full pricing model (Black-Scholes or a more sophisticated stochastic volatility model) running across thousands of contracts. Firms increasingly use GPUs for parallel pricing computation and FPGAs for the most latency-sensitive calculations.

Monitoring and kill switches. Automated systems can go catastrophically wrong. A bug in a quoting algorithm can send millions of erroneous orders in seconds. Every market making firm has kill switches that can pull all quotes from all exchanges within milliseconds, circuit breakers that detect abnormal P&L or position changes, and real-time dashboards that human risk managers monitor throughout the trading day. The Knight Capital incident of 2012 - where a software deployment error caused $440 million in losses in 45 minutes - remains the cautionary tale for the industry.

---

Top Market Making Firms

A small number of firms dominate global market making. They combine quantitative research, engineering talent, and massive capital deployment to provide liquidity across asset classes and geographies.

Citadel Securities is the largest market maker in the world by volume, handling roughly 25% of all US equity volume and a significant share of options and fixed income markets. Founded by Ken Griffin and operating independently from Citadel's hedge fund, the firm has invested heavily in technology and talent. Headquartered in Miami with major offices globally.

Virtu Financial is a publicly traded market maker known for its technology-driven approach. Virtu famously reported only one losing trading day in over 1,200 between 2009 and 2014, illustrating the consistency of well-executed market making. The firm makes markets across equities, fixed income, currencies, and commodities on over 200 venues globally.

Optiver is a Dutch proprietary trading firm founded in 1986 in Amsterdam, specialising in options and ETF market making. Optiver is known for its quantitative culture and technology investment, operating from Amsterdam, Chicago, Sydney, and Shanghai. See our Optiver guide for more on the firm.

Jump Trading is a Chicago-based firm that operates at the frontier of speed and technology. Jump is a major market maker in futures, equities, and crypto, and has invested in infrastructure ranging from microwave networks to custom silicon. They're known for recruiting top engineering and research talent.

Jane Street is a New York-based quantitative trading firm that is one of the largest ETF and options market makers globally. Jane Street's approach is research-driven, with a culture rooted in OCaml programming and rigorous quantitative thinking. The firm trades across equities, bonds, options, ETFs, and commodities. Read more in our Jane Street interview guide.

For a broader overview of firms in this space, including interview processes and career paths, see our guide to proprietary trading firms.

---

Frequently Asked Questions

How does a market making strategy differ from other trading strategies?

A market making strategy profits from providing liquidity - earning the bid-ask spread by continuously quoting both a buy and a sell price. Most other trading strategies are directional: they profit from predicting whether prices will rise or fall. A trend-following strategy buys assets that are going up. A value strategy buys assets that are underpriced. A market maker is agnostic about direction - they're happy to buy or sell, as long as they can do both and capture the spread between. The risk profile is also different. Directional strategies have concentrated exposure to market moves. Market makers have distributed exposure across many small positions, with the main risk being inventory accumulation during one-sided flow.

How do market making algorithms work?

A market making algorithm continuously calculates a fair value for an instrument, places bid and ask orders at a spread around that fair value, monitors incoming fills and adjusts inventory, skews quotes to manage risk, and updates everything when market conditions change. The algorithm ingests real-time market data (prices, order book depth, trades), runs it through a fair value model that might incorporate signals from correlated instruments, and outputs bid and ask prices. When a trade executes, the algorithm updates its inventory position and adjusts subsequent quotes to encourage flow that reduces exposure. Modern market making algorithms process thousands of data updates per second and regenerate quotes within microseconds. They also incorporate risk limits, position constraints, and kill-switch logic to prevent runaway losses.

Is market making profitable in 2026?

Market making remains profitable in 2026, but profit margins have compressed significantly over the past two decades. Spreads have narrowed as competition has intensified and technology has improved. The firms that remain profitable are those with scale (quoting across thousands of instruments and multiple asset classes), superior technology (lower latency, better pricing models), and efficient risk management. Smaller or less technologically sophisticated market makers have been squeezed out. The barrier to entry is high - a new entrant would need to invest tens of millions in technology and infrastructure before earning a single pound. For the firms that can compete, however, market making generates consistent returns with relatively low variance compared to directional strategies.

What is automated market making in DeFi?

Automated market making (AMM) in decentralised finance replaces the traditional order book with a mathematical formula that determines prices based on the ratio of assets in a liquidity pool. The most common formula is the constant product model (x * y = k), used by Uniswap and its forks. Liquidity providers deposit pairs of tokens into a pool and earn trading fees when others swap against the pool. The price adjusts automatically as the ratio of tokens changes. AMMs differ from traditional market making in several important ways: there's no order book, prices are set by formula rather than by individual quotes, anyone can be a liquidity provider (no need for expensive infrastructure), and the primary risk is impermanent loss rather than traditional inventory risk. Impermanent loss occurs when the price ratio of the deposited tokens changes, leaving the liquidity provider worse off than if they'd simply held the tokens.

What qualifications do you need to work at a market making firm?

Most market making firms hire people with degrees in mathematics, physics, computer science, statistics, or engineering. For trading and quantitative research roles, strong probability and statistics skills are essential, along with programming ability in Python, C++, or both. For technology roles, systems programming, low-latency networking, and FPGA development are valued. Some firms - particularly Optiver, Jane Street, and IMC Trading - hire talented undergraduates directly, while others prefer candidates with postgraduate qualifications. Beyond formal credentials, firms test for quick quantitative reasoning, the ability to make decisions under uncertainty, and intellectual curiosity. Interview processes typically involve probability puzzles, coding challenges, and market-making simulations where candidates must quote prices and manage a mock portfolio in real time.