Your Broker Is Betting Against You: How CFD Market Makers Profit From Your Losses

This is not a fringe practice or a regulatory gray area. It is the standard operating model of the retail CFD industry, disclosed in the legal documentation of virtually every major broker, acknowledged by regulators, and well understood by anyone who has worked inside a CFD brokerage. The question is not whether it happens. The question is what it means for your trading, and whether a better alternative exists.

The A-Book Model: What Fair Execution Looks Like

To understand the B-book model, it helps to first understand the A-book model, which is how most people assume their broker operates. In an A-book (also called STP, or Straight-Through Processing) arrangement, the broker forwards your order to an external liquidity provider. This might be a major bank, a prime broker, or an electronic communication network that aggregates liquidity from multiple sources.

In this model, the broker's revenue comes exclusively from the spread markup or commission charged on each trade. The broker adds a small increment to the liquidity provider's spread and passes the rest through. If the liquidity provider is quoting EUR/USD at 1.1000/1.10008, the A-book broker might quote you 1.09995/1.10013, pocketing the 0.5-pip difference on each side. The broker has no exposure to the outcome of your trade. If you win, the liquidity provider pays. If you lose, the liquidity provider profits. The broker earns the same markup either way.

True A-book execution creates genuine neutrality. The broker's financial interest is aligned with yours in the sense that the broker wants you to trade more and trade for longer. A winning client is a client who stays, trades more volume, and generates more spread revenue over time. An A-book broker benefits from your success because your success means continued business.

However, pure A-book brokers are rare in the retail CFD space. The reason is economic. An A-book broker might earn $5 to $15 per standard lot in spread markup. To generate meaningful revenue, it needs very high trading volumes, which requires competitive pricing, which compresses margins further. The A-book model works as a business, but it is far less profitable per client than the alternative, which is why most retail brokers have adopted the B-book model instead.

The B-Book Model: When Your Broker Becomes Your Counterparty

In a B-book model, the broker internalizes your trade. There is no external liquidity provider. The broker itself is the buyer when you sell and the seller when you buy. Your profit and loss is directly and inversely correlated with the broker's profit and loss.

The economics of B-booking are extraordinarily attractive from the broker's perspective. ESMA data consistently shows that between 74% and 89% of retail CFD accounts lose money. If a broker internalizes all retail order flow, the expected value of client losses is massively positive. The broker earns spread revenue from every trade (both winners and losers), earns overnight financing fees from every position held overnight, and then captures the net trading losses of its client base. When three-quarters or more of your clients lose money, warehousing their trades is not a risk. It is a near-guaranteed profit.

Consider the economics in concrete terms. A mid-sized CFD broker with 5,000 active retail clients, each averaging $150,000 in monthly notional volume, generates approximately 37,500 standard lots per month. At $20 per lot in spread revenue, that produces $750,000 monthly in spreads alone. If 78% of these clients are net losers with an average monthly loss of $500, client losses total $1,950,000. The 22% of winning clients with average monthly gains of $350 cost the broker $385,000. The net from B-booking is $1,950,000 minus $385,000 plus $750,000, totaling $2,315,000 per month. Annual revenue from this model exceeds $27 million. This is why the retail CFD industry spends hundreds of millions on marketing and advertising. Every new depositing client has a high probability of being net profitable for the broker.

The Hybrid Model: Profiling, Sorting, and Adverse Treatment

In practice, most large CFD brokers operate a hybrid model. They do not B-book every client indiscriminately. Instead, they use sophisticated analytics to categorize clients by expected profitability and route flow accordingly.

New clients with small deposits and no trading history are typically B-booked by default. The probability that these clients will be net losers is very high, and the broker wants full exposure to their expected losses. As clients trade and generate a track record, the broker's risk management system monitors their performance.

Clients who demonstrate consistent profitability over a 30 to 90 day window are flagged. The broker then has several options for managing its exposure to these winning traders. The most common approach is to A-book the profitable client's trades going forward, routing them to an external liquidity provider so the broker no longer has counterparty exposure. The broker still earns the spread markup but avoids paying out the client's gains from its own book.

Other documented approaches are more aggressive and more concerning. Some brokers implement differential pricing for profitable traders, widening spreads selectively so the cost of trading increases and the trader's edge is eroded. Some brokers introduce execution delays for profitable traders, adding latency between order submission and execution. During the delay, the broker can observe price movement and decide whether to fill at the original price or reject and reprice the order. This is known as last-look execution, a practice that has faced increasing regulatory scrutiny.

And in some documented cases, brokers simply close the accounts of consistently profitable traders. The broker invokes broad terms-of-service clauses to terminate the relationship, often without explanation. The trader is paid their current balance and told their account is no longer welcome. This practice has been reported extensively in trading forums, with some brokers developing reputations for banning winners.

Documented Regulatory Actions and Industry Cases

The conflict of interest inherent in CFD market-making is not theoretical. It has been documented in regulatory enforcement actions across multiple jurisdictions.

In 2014, the FCA imposed significant fines on FXCM, at the time one of the world's largest retail forex and CFD brokers. The FCA found that FXCM had failed to pass on favorable price improvements to clients. When FXCM's liquidity provider offered a better price than the client's requested price, FXCM kept the improvement as additional profit rather than passing it to the client. Conversely, when prices moved against the client, the full unfavorable slippage was passed through. This systematic asymmetric treatment of price improvements meant the execution was structurally biased against clients. FXCM was later forced to exit the US market entirely following additional regulatory action by the CFTC for similar issues.

In 2017, the Cyprus Securities and Exchange Commission (CySEC) imposed fines on several CFD brokers for failures in order execution and conflict of interest management. Cyprus is the regulatory home of many EU-licensed CFD brokers, and CySEC's enforcement actions have targeted practices including artificial price manipulation, failure to execute at best available prices, and inadequate disclosure of the B-book model to clients.

In 2020, ASIC took its most significant action against the CFD industry. Following a comprehensive review, ASIC implemented product intervention orders that restricted leverage, banned incentive schemes, and required standardized risk warnings. In its decision rationale, ASIC explicitly stated that the market-making model created inherent conflicts of interest that could not be adequately managed through disclosure alone. ASIC found that many brokers operated business models that were commercially dependent on client losses and that this dependence created incentives that were fundamentally adverse to client interests.

Beyond formal regulatory actions, a substantial body of evidence from former industry employees describes internal practices designed to maximize client losses. These include targeting marketing at demographics with low financial literacy, designing platform features and notifications that encourage overtrading, using behavioral analytics to identify clients likely to deposit more after losses, and implementing execution policies that systematically disadvantage clients. Individual accounts should be evaluated critically, but the consistency of these reports across multiple brokers, countries, and time periods forms a coherent pattern.

Why Profitable Traders Get Banned

The B-book model creates a perverse situation where a broker's best clients, from a trading skill perspective, are actually the broker's worst clients from a business perspective. A consistently profitable trader extracts money from the broker's book. Every dollar the trader makes is a dollar the broker loses.

In an A-book model, profitable traders are desirable. They trade more, they stay longer, and they generate more commission revenue over their lifetime. In a B-book model, profitable traders are a liability. They reduce the broker's net capture rate and, if they trade significant volume, can materially impact the broker's profitability.

This is why account closures of profitable traders occur. The broker's risk management system identifies clients who are consistently extracting money from the B-book. The broker faces a choice: A-book the client's trades going forward, accepting reduced revenue, or terminate the relationship entirely. Many brokers choose termination, calculating that the client's account is not worth maintaining even on an A-book basis because the tight spreads required to retain a sophisticated trader would further compress margins.

The result is a system that rewards failure and punishes success. Losing traders are welcome. They are the product. Winning traders are a problem to be managed, degraded, or removed. This is not a conspiracy. It is the rational economic behavior of a business whose primary revenue is derived from client losses.

The DEX Orderbook Model: Neutral Infrastructure

Decentralized exchanges like Hyperliquid operate on a model that eliminates the structural conflicts of CFD brokers. The exchange is not a counterparty. It does not take the other side of trades. It does not maintain a B-book. It does not profile clients by profitability. It does not degrade execution for winners.

On Hyperliquid, the exchange operates a fully on-chain order book. When you place an order, it enters a public, transparent matching engine. Your order is matched against another trader's order. The only entities on the other side of your trade are other market participants: other traders, market makers, and arbitrageurs who have chosen to take the opposite position.

The exchange earns a small fee from both the maker and the taker on every trade. This fee is identical whether the trader wins or loses, whether the trader is new or experienced, whether the trader is profitable or unprofitable. The exchange has zero financial interest in the outcome of any individual trade or in the aggregate profitability of its user base. Revenue is a function of volume, not of client losses.

This creates a fundamentally different incentive structure. The exchange is incentivized to attract and retain all traders. Winning traders are as valuable as losing traders because both generate volume. The exchange is incentivized to offer the tightest possible spreads, the fastest possible execution, and the most reliable possible infrastructure, because these factors drive volume. There is no incentive to widen spreads, delay execution, or ban profitable participants.

The order book is public and verifiable. Every order, fill, cancellation, and liquidation is recorded on the blockchain. You can audit your own execution by comparing your fill price to the order book state at the block timestamp. Asymmetric slippage is not a theoretical impossibility but a mathematical one when the matching engine follows deterministic, publicly auditable rules. There is no last-look execution because there is no entity with the unilateral power to reject or reprice orders after submission.

Price discovery is genuine and market-driven. Prices on Hyperliquid reflect the actual intersection of supply and demand on the order book, kept in line with global markets by arbitrageurs and the funding rate mechanism. There is no risk desk adjusting quotes. There is no proprietary price feed. There is no spread widening at the discretion of a counterparty. The rules are encoded in smart contracts, applied uniformly, and transparent to everyone.

What This Means for Your Trading

If you are currently trading CFDs, you are operating in a system with a structural conflict of interest. Your broker may be professional, regulated, and well-intentioned. But the architecture of the product means that your losses benefit the broker and your profits cost the broker. No amount of compliance infrastructure or regulatory oversight can fully neutralize an incentive that is embedded in the product's design.

For traders who are still developing their skills, the conflict may seem academic. When you are losing money, the mechanism through which you lose matters less than the fact that you are losing. But for traders who are developing a genuine edge, the implications are significant. A strategy with a positive expected value can be degraded or destroyed by adverse execution, selective spread widening, or account closure. On a B-book CFD platform, success itself becomes a risk factor.

On a decentralized exchange like Hyperliquid, none of these dynamics exist. Your execution quality is constant. Your fees are constant. Your access to the platform is constant. The protocol treats every participant identically because it is architecturally incapable of discrimination. Equal treatment is not a policy decision that can be reversed. It is a property of the code.

The shift from CFD brokers to decentralized perpetual futures exchanges represents more than a technological change. It is a structural realignment of who benefits from the trading process. For the first time, retail traders can access leveraged derivative markets on infrastructure that is genuinely neutral, transparently governed, and fundamentally incapable of profiting from their losses. The broker is no longer at the table. The house does not have an edge. The only thing that determines your outcome is your trading.

Hyperdash Tip

When evaluating any trading venue, ask one critical question: does this platform make more money when I lose? If the answer is yes, or if the answer is unclear, you have a structural conflict of interest working against you. On Hyperliquid, the answer is definitively no. The protocol earns the same fee regardless of your outcome. Use Hyperdash to trade on Hyperliquid with a professional-grade interface that includes advanced order types, real-time portfolio analytics, multi-chart layouts, and comprehensive market data, all running on infrastructure that has zero interest in whether your trades succeed or fail. Your edge should be the only thing that matters. On Hyperliquid, it is.

Frequently Asked Questions

Is the B-book model illegal?

No. B-booking is a legal and regulated practice in all jurisdictions where CFDs are permitted. Regulators require brokers to disclose their execution model, manage conflicts of interest, and comply with best execution obligations. However, the fact that a practice is legal does not mean it serves the client's interest. The B-book model creates a structural conflict that regulators have acknowledged and attempted to mitigate through rules, disclosures, and product restrictions. The fundamental misalignment, that the broker profits from client losses, remains an inherent feature of the model regardless of regulatory compliance.

How can I determine if my CFD broker uses a B-book?

Check the broker's order execution policy and conflicts of interest disclosure, both of which regulated brokers are required to publish. Look for language stating that the broker 'may act as principal' or 'as counterparty' to client transactions, or that the broker acts as a 'market maker.' If any of these phrases appear, the broker operates a B-book for at least a portion of its client flow. Most major retail CFD brokers use a B-book or hybrid model. Brokers that operate a pure A-book model typically market this fact prominently, as it is a competitive differentiator.

Do decentralized exchanges have conflicts of interest?

DEXs have substantially fewer conflicts than B-book CFD brokers, though they are not entirely conflict-free. The development team or governance token holders may have interests that influence protocol decisions. On AMM-based DEXs, liquidity providers take the opposite side of trades, though this does not apply to order book exchanges like Hyperliquid. The critical structural difference is that no single entity on a DEX has the ability to manipulate prices, widen spreads, degrade execution for specific traders, or profit from individual client losses. The rules are encoded in public, auditable smart contracts that apply uniformly to all participants. This architectural neutrality is the core advantage over the CFD model.

What happens to profitable traders on Hyperliquid?

Nothing changes. Your execution quality, your fees, your platform access, and your withdrawal capability remain identical regardless of your profitability. There is no mechanism to classify traders by outcome, no system to route profitable flow differently, and no ability to restrict or terminate accounts based on trading performance. The protocol is programmatically incapable of treating participants differently based on their results. This is perhaps the single most important distinction between decentralized exchange infrastructure and the B-book CFD model: equal treatment is not a corporate policy that can be changed. It is an immutable property of the protocol's design.