Every quote on your screen looks like a fact. It is actually a composite: a price assembled by your broker from streams supplied by competing liquidity providers, shaped by an execution model the marketing page rarely explains. This sheet opens that machinery — aggregation, depth, who takes the other side of your trade and when — and ends with the seven questions that make any broker's execution checkable, including ours.
There is no single price: aggregation from liquidity providers
Forex has no central exchange and therefore no official tape. The market is a mesh of banks, non-bank market makers, and electronic venues, each quoting prices at which it is willing to deal. A retail broker connects to several of these — its liquidity providers — and each streams a continuous feed of prices: a bid at which it will buy from the broker and an ask at which it will sell.
The broker's aggregation engine merges those streams and shows you the best available combination: the highest bid any provider will pay, the lowest ask any will accept. The EUR/USD quote on your platform is that composite, refreshed many times a second, plus whatever markup your broker adds. This is why two reputable brokers show fractionally different prices at the same instant, and why neither is lying: each is honestly reporting the best of the streams it receives. “The price” is a convenient fiction; what exists is a crowd of prices, and your broker's job is to stand in front of it.
Aggregation quality is one of the real differences between brokers. More providers competing in the engine generally means tighter spreads and steadier liquidity when conditions turn fast — and a broker proud of its aggregation can usually say how many providers feed it. Keep that thought; it becomes question two of seven.
The order book you don't see: depth, top-of-book, and why size moves price
The quote on your screen is the top of a book — the best price, available for a limited amount. Behind it sit further levels: more size, at progressively worse prices. A platform shows you top-of-book by default, which quietly hides the most useful fact about it: the best price is only firm for the size attached to it.
Work the arithmetic. Suppose the aggregated book bids 1.0850 for 5 standard lots, and 1.0849.5 — half a pip lower — for the next 5. You sell 10 lots at market: the first 5 fill at 1.0850, the rest at 1.08495. Your average price is 1.084975, a quarter-pip below the quote you clicked — $25 across the position at $10 per pip per lot. Nobody moved the price against you; your order simply consumed the first level and dealt at the second. A 10-lot buy walks the ask side upward with the same arithmetic in reverse.
Retail-sized orders in major pairs almost never exhaust a level in normal hours. But depth is not constant: around news and at thin hours, providers pull quotes and the levels behind the best price grow sparse — which is when modest orders start walking the book, and when slippage statistics earn their keep. Depth, not the headline spread, is what actually fails first in fast markets. The headline spread is what brokers advertise; depth is what you trade against.
One more piece of machinery deserves daylight: last look. Many liquidity providers reserve a final moment, after receiving an order, to accept or decline it at the quoted price — a practice tolerated in FX precisely because quotes are streamed blind to thousands of recipients at once. Used as designed, it protects providers from dealing on stale prices; used aggressively, it lets them reject exactly the trades that would have favored you. It is one structural reason a streamed price is an invitation rather than a promise, and why rejection statistics — question four below — are worth asking for.
Execution models without euphemism: who takes the other side
Every trade has a counterparty. The execution-model question is simply: who is yours? Strip the marketing vocabulary and there are two pure answers and one honest one.
Market making — B-book, in desk language: the broker itself takes the other side. You buy, it sells to you, and your positions sit on its risk book. This is neither new nor inherently abusive — exchanges have always had market makers, and internalizing flow lets a broker quote tight spreads in sizes the external market would not honor. The structural fact remains: on internalized flow, your loss is the dealing desk's gain until the desk hedges its net exposure.
Agency execution — A-book, often labeled STP: the broker passes your order through to its liquidity providers and earns a markup or commission on the way. The provider, not the broker, holds the other side. The broker's profit is your transaction cost, not your loss, which aligns incentives better — but pure agency brokers pass through whatever the external market actually offers, including its requotes, its thinner depth at speed, and its news-time spreads.
The hybrid reality: most retail brokers run both books at once, routing flow by rules you cannot see — by client profile, instrument, size, or market conditions. “We are not a market maker” is, at many firms, a statement about one book of two. The honest questions are therefore not “which model are you?” but “when is my order internalized, when is it passed through, and who decides?” Any broker can answer that. Most are rarely asked.
Conflicts of interest: where they live, and what regulation actually does
Name the conflicts plainly. On an internalized trade, the broker's dealing revenue rises when your trade loses. A broker controls the markup inside its own spread, and widening it is invisible unless you measure against a benchmark feed. Execution timing, requote policy, and stop-trigger sources are all set by the party that may also be your counterparty. None of this is hidden wrongdoing — it is the architecture of the product, and it is why this article exists.
Honesty also requires the other half of the ledger. Internalization is not automatically against you: a desk netting thousands of client orders can fill them tighter and faster than the external market would, and it absorbs sizes and hours that outside providers decline. Plenty of firms run hybrid books for operational reasons rather than predatory ones. The point of naming the conflicts is not to declare every broker guilty — it is that you cannot distinguish a well-run hybrid from a badly run one by reading its homepage. You distinguish them by what they disclose and what you measure.
Regulation constrains the architecture rather than abolishing it. Licensed brokers in serious jurisdictions owe clients best execution and must publish an order-execution policy describing how orders are handled and where conflicts sit. Client money must be segregated from the firm's own. Regulators require firms to monitor execution quality, and several publish or compel statistics on it. What regulation does not do is watch your individual fills, audit your statement, or promise that the published policy matches lived execution minute by minute. That last step has no institution behind it. It has you.
The formal version of this sheet — the document a regulator holds us to.
The checklist: seven execution questions for any broker — including us
Send these to any broker you are evaluating. Every one has a factual answer; none can be honestly answered with an adjective.
- Who takes the other side of my trades — your own desk, external liquidity providers, or both? Under what rules is an order internalized rather than passed through?
- How many liquidity providers feed your pricing, and does the count differ between marketing and the order-execution policy?
- What were your average execution speed and your slippage statistics over the last quarter — positive and negative, separately?
- What share of orders was requoted or rejected in that period, and under what conditions does that happen?
- How is your spread constructed — raw feed plus a stated markup, or a managed spread? What exactly is the markup on the account type I would hold?
- What price source triggers stop orders, and how are stops and limits handled through news and gaps? (This sheet's companion on slippage explains why that wording matters.)
- Where is your order-execution policy published, when was it last revised, and what changed?
Grade the replies on form before content. Numbers, document links, and named conditions are checkable claims; “deep liquidity” and “lightning execution” are decoration. A broker that answers question three with two distributions and a date is telling you something real about its operations — whatever the numbers say. A broker that cannot answer it at all has answered it anyway.
Every broker says trust us. We'd rather you check.
Send the checklist as written. Answers in numbers, in writing — that is the standard we are asking you to hold everyone to.
Reading an execution statement like an auditor
The answers above are claims; your account statement is evidence. Auditing it takes four columns and twenty minutes. Export your filled orders and line up: requested price, executed price, the difference signed in pips, and the timestamp to the second. Then read like an auditor rather than a customer — in aggregates, not anecdotes.
Three aggregates do most of the work. The slippage distribution: mostly zero, with small tails in both directions — a tail that only ever points against you is the single loudest red flag in retail execution. The conditional pattern: worse fills clustered at news minutes and thin hours are physics; worse fills scattered through calm markets are a question. And the comparison: your measured experience set against the broker's published statistics from question three. Where the two diverge, you have something better than a complaint — you have data, and a specific, answerable letter to write.
This is the quiet point of the whole sheet. Aggregation, depth, execution models, and conflicts are not things a retail trader can change — but every one of them leaves fingerprints on a statement you already own. A broker chosen on measured fills, published policies, and answered questions is a decision; one chosen on a spread banner is a guess.