A-Book vs B-Book Broker: How Forex Models Really Work

A-book and B-book describe how a forex broker handles the risk on your trade, not how fast your order fills. An A-book broker passes your trade to the real market and earns a small markup or commission. A B-book broker keeps your trade in-house and becomes the other side of it, so when you lose, the broker keeps the difference. Most brokers today actually run both at once, splitting clients between the two based on risk profile. That third setup is called a hybrid model, and by 2026 it's become the industry default rather than the exception.

If you're a trader trying to work out which type of broker you're dealing with, or a broker trying to decide how to structure your own risk book, this article gives you the plain answer, backed by real numbers, without the sales pitch either side usually attaches to it.

The quick answer

• A-book: your broker sends your order to a liquidity provider (a bank or non-bank market maker) and makes money from the spread or a commission, regardless of whether you win or lose.
  

• B-book: your broker keeps your order on its own books and makes money when you lose. Your win is the broker's loss.
  

• Hybrid: the broker routes some clients (usually profitable or high-volume ones) to A-book and keeps the rest, typically smaller or newer accounts, in B-book.

None of these three is automatically a scam. All three are used by fully regulated, licensed brokers around the world. The difference is what each one is built for.

How an A-book broker actually makes money?

When you place a trade with an A-book broker, your order goes out to a liquidity provider, often a Tier 1 bank, a prime broker, or a non-bank market maker with deep pricing. The broker isn't betting against you. It earns its keep from the gap between the price it gets from the liquidity provider and the price it gives you, plus sometimes a flat commission per lot.

Here's a simple example. Say a liquidity provider quotes EUR/USD at 1.0850. The broker shows you 1.0852. That two-pip gap is the broker's revenue on that trade, whether you end up winning or losing your position. Your outcome and the broker's revenue are disconnected. That's the core appeal of A-book: no conflict of interest between what's good for you and what's good for the broker.

The tradeoff is thinner margins. A-book brokers need real trading volume to turn a profit, which is why they tend to court higher-net-worth clients, professional traders, and institutional flow rather than small retail deposits.

How a B-book broker actually makes money?

A B-book broker doesn't send your order anywhere. It becomes your counterparty directly. If your trade loses, the broker keeps that loss as revenue. If your trade wins, the broker pays you out of its own book, essentially taking on the risk itself instead of passing it to the market.

This is the model most new traders find uncomfortable once they understand it, because it means the broker's short-term financial interest and the client's are pointed in opposite directions. But it's worth being precise here: B-book on its own isn't illegal or inherently predatory. Every major regulator, the FCA, ASIC, CySEC, and others, permits it, provided the broker discloses it, manages the risk properly, and doesn't manipulate execution to force losses.

The reason B-book exists at all is simple. Most retail trading accounts lose money over time. That's not opinion, it's the actual disclosed data. Regulators in Europe require brokers to publish the percentage of retail clients who lose money trading CFDs, and that mandated figure has consistently landed between 74 percent and 89 percent under ESMA's methodology. In the UK, an April 2026 rolling index of 14 FCA-authorised CFD brokers put the average loss rate at 69.9 percent, with individual brokers ranging from 51 percent to 82 percent. In the US, the CFTC has reported similar patterns, generally in the 70 to 80 percent range. When the majority of clients are statistically likely to lose, a broker that keeps that flow internally rather than paying a liquidity provider to hedge it can be more profitable, as long as it manages the concentration risk correctly.

I've sat in enough conversations with broker founders to know the B-book question always gets asked the wrong way. It's never really "should I B-book," it's "do I actually have the risk controls to B-book safely." A broker running B-book without exposure limits, real-time monitoring, and a capital buffer isn't running a business model, it's running a bet. The brokers that get into trouble aren't the ones who B-book. They're the ones who B-book blind.

Why hybrid has become the default in 2026?

Pure A-book caps your margins. Pure B-book concentrates your risk in one direction, and if enough clients win at once (which does happen during strong trending markets or news events), it can hurt the broker badly. That's why almost every established retail broker now runs a hybrid book, splitting order flow based on a client's trading pattern rather than treating every client the same way.

A typical hybrid setup uses a risk engine that scores each client on signals like account size, trade frequency, holding time, and historical profitability. Scalpers and consistently profitable traders often get routed to A-book, because internalising their flow is genuinely risky for the broker. Casual or newer traders, statistically more likely to lose, are more often kept in B-book. This client classification approach is now common enough across the industry that it's less a competitive edge and more a baseline requirement for running a sustainable brokerage.

What's changed more recently is the technology behind that decision. A growing number of brokers now use machine-learning-driven risk engines that reclassify clients in near real time rather than on a fixed schedule, adjusting exposure the moment a trader's behaviour shifts. It's a natural extension of the same AI-driven infrastructure that's reshaping execution and pricing across the industry more broadly.

A-Book vs B-Book vs Hybrid: Side-by-Side Comparison

A-Book Broker

• Who takes the market risk: Liquidity provider

• Broker's revenue source: Spread markup or commission

• Conflict of interest with client: None

• Typical broker margin: Thin and volume-dependent

• Best suited for: High-volume and professional trading flow
  

B-Book Broker

• Who takes the market risk: The broker itself

• Broker's revenue source: Client losses

• Conflict of interest with client: Yes, by design

• Typical broker margin: Wider, but carries more risk

• Best suited for: Small, low-frequency retail trading flow
  

Hybrid Broker

• Who takes the market risk: Split between the broker and liquidity providers based on the client profile

• Broker's revenue source: A combination of spread/commission revenue and internalised client flow

• Conflict of interest with client: Only on the B-booked portion

• Typical broker margin: More balanced

• Best suited for: Most modern retail brokerages
  

What this actually means if you're a trader?

Knowing whether your broker A-books or B-books your flow doesn't tell you whether you'll win or lose. It tells you where the incentive sits. A regulated broker, properly supervised, has strong reasons (license conditions, audits, reputational risk) not to manipulate execution even when it's technically the counterparty to your trade. An unregulated one running B-book with no oversight is a different story entirely, and that distinction matters far more than the book model itself.

If you want to get a sense of which model a broker leans on, a few practical signs help. Brokers that are pure market makers (no external liquidity, in-house dealing desk only) are almost always running mostly B-book. Brokers that advertise "STP" or "ECN" execution are typically leaning A-book, though as we cover in our related guide on execution models, that label describes how your order is filled, not automatically where the risk ends up. And brokers that publish their mandated loss-rate disclosure (as FCA and ESMA-regulated firms must) are giving you a genuine, audited data point rather than a marketing claim.

What this means if you're building a brokerage?

If you're setting up a new brokerage, the honest starting point is this: almost nobody launches pure A-book and survives on margin alone unless they're targeting institutional or high-volume flow specifically. And almost nobody should run pure B-book without a properly built risk engine, because a handful of correctly-timed trades during a volatile week can wipe out months of B-book profit in a single session.

The founders who come to us already asking about hybrid architecture, rather than "should I B-book my clients," are usually the ones who've done their homework. The ones who haven't tend to assume B-book means guaranteed profit. It doesn't. It means managed risk, and managed risk still requires real infrastructure, real capital buffers, and real oversight.

Getting the classification logic, the risk engine, and the regulatory disclosure right from day one is a technology and compliance decision as much as a business one, which is exactly where a lot of new brokerages underinvest early and pay for it later.

Frequently asked questions

Is B-book trading illegal?
No. B-book is legal and widely used by regulated brokers worldwide, including many licensed under the FCA, ASIC, and CySEC. It becomes a problem only when a broker fails to disclose it, manipulates execution, or runs it without proper risk controls.

Which is better for traders, A-book or B-book?
Neither is inherently better. What matters more is whether the broker is properly regulated, discloses its model and loss rates, and doesn't interfere with your execution. A well-run B-book broker with fair pricing can be perfectly safe. A poorly run A-book broker can still have hidden costs through slippage or requotes.

How do I know if my broker is A-book or B-book?
Most brokers won't state it outright, but you can look at their regulatory disclosures, ask directly whether they hedge client flow externally, and check whether they publish a mandated loss-rate percentage. A licensed order execution policy document, which regulated brokers must provide, usually spells this out.

Can a broker be both A-book and B-book at the same time?
Yes, this is the hybrid model, and it's now the standard approach for most retail forex and CFD brokers rather than the exception.

Does B-book mean the broker is trading against me on purpose to make me lose? Not in a well-regulated setup. The broker's revenue improves when you lose, but that's a structural incentive, not proof of manipulation. Regulators require fair execution regardless of which book a trade sits in, and manipulating prices or fills to force losses is a serious compliance violation that can cost a broker its license.

How hybrid routing works in practice

The hybrid model sounds abstract until you see where it lives: in routing rules on the bridge, reviewed daily by whoever runs risk. A typical rule set segments flow by client profile, instrument and current exposure. New or unprofitable-flow clients may be internalised; consistently profitable or large clients get routed to external liquidity; and exposure that accumulates beyond a threshold in any instrument gets hedged regardless of which book the originating client sits in. The mechanics of that routing layer are covered in our liquidity bridge guide.

What matters operationally is that these are monitored rules, not set-and-forget configuration. Exposure concentrations, execution quality and client profitability distributions change, and the routing has to change with them. That monitoring is exactly what risk tooling such as RiskMan automates, and what a managed 24/7 dealing desk does as a service — the software and the staffed service are separate offerings that solve the same underlying problem at different levels of involvement.

What regulators expect regardless of model

Neither model is prohibited in the major regulatory regimes, but both come with conduct obligations. Regulated brokers are expected to disclose their execution model in their order execution policy, publish client loss-rate statistics where the regime requires it, and demonstrate best execution regardless of which book a trade sits in. Conflicts of interest inherent in B-booking must be identified and managed, not merely disclosed. The practical implication for anyone choosing a model: the compliance work is similar in both cases, and the choice between A-book, B-book and hybrid should be made on capital, risk appetite and operational capability — not on the assumption that one model avoids regulatory attention.

Disclaimer: This article is for general educational and informational purposes only. It does not constitute financial, investment, legal, or regulatory advice, and it isn't a recommendation to use any particular broker or business model. Loss-rate statistics cited here come from publicly available regulatory disclosures (ESMA, FCA, CFTC) current as of 2026 and can change as regulators update their methodology. If you're a trader, consider your own risk tolerance and consult a licensed financial adviser before making trading decisions. If you're building or operating a brokerage, consult a licensed compliance professional and your relevant regulator before choosing or changing a risk model, since requirements vary significantly by jurisdiction.