A-book, B-book, and the clues in your fills
Posted: Thu Oct 29, 2026 11:08 am
Brokers handle client orders in different ways. Some pass them on to the market, often called A-book. Others keep the other side of the trade internally, known as B-book, profiting when clients lose. Many use a hybrid, routing some flow one way and some the other.
A hybrid model creates a potential conflict of interest, since the broker benefits from client losses on the internalised part. Reputable firms manage this transparently, and regulators require disclosure of the order execution policy.
Traders cannot see inside the process, but some patterns in execution may offer hints. Consistent one-sided slippage, delays on profitable trades, differences in fill quality between profitable and unprofitable periods, and sudden changes after a run of success can all be clues.
None of these proves anything alone. They could reflect market conditions or technical issues. A larger dataset is needed to separate noise from pattern.
The practical response is documentation. A clean log of fills over time makes it possible to raise a concern with evidence, or to decide to move to a different broker.
Some traders open two accounts at the same broker with different profiles, such as one trading actively and one trading rarely, and compare the fill quality. A consistent difference suggests that the broker's handling depends on the account's behaviour. It is not conclusive, but it offers a structured way to test the question.
What does your own execution data suggest about how your orders are handled?
A hybrid model creates a potential conflict of interest, since the broker benefits from client losses on the internalised part. Reputable firms manage this transparently, and regulators require disclosure of the order execution policy.
Traders cannot see inside the process, but some patterns in execution may offer hints. Consistent one-sided slippage, delays on profitable trades, differences in fill quality between profitable and unprofitable periods, and sudden changes after a run of success can all be clues.
None of these proves anything alone. They could reflect market conditions or technical issues. A larger dataset is needed to separate noise from pattern.
The practical response is documentation. A clean log of fills over time makes it possible to raise a concern with evidence, or to decide to move to a different broker.
Some traders open two accounts at the same broker with different profiles, such as one trading actively and one trading rarely, and compare the fill quality. A consistent difference suggests that the broker's handling depends on the account's behaviour. It is not conclusive, but it offers a structured way to test the question.
What does your own execution data suggest about how your orders are handled?