ECN, STP and Market Maker Brokers: How Execution Models Really Work

ECN, STP and market maker describe what your broker does with your order after you click. The labels are used loosely in marketing, and the honest answer is that execution quality matters far more than the acronym on the account page.

ECN, STP and Market Maker Brokers at a glance

What it is and why it works

Every retail order has to end up somewhere. The execution model describes where. In the market maker or dealing-desk model, the broker becomes your counterparty: you buy, it sells, and the position now sits on the broker’s own book. It manages that exposure by netting clients against each other, your long offsets someone else’s short, and hedging whatever imbalance is left in the interbank market, or choosing not to hedge it at all.

In the STP model the broker does not warehouse the risk. It routes your order straight through to one or more liquidity providers, typically banks and non-bank market makers, and earns from a markup applied to the price it receives or from an explicit commission. In the ECN model orders enter a shared electronic pool where many participants post bids and offers, so you can in principle see depth of market and your order can be filled by another participant rather than by a designated dealer. ECN pricing is usually raw, the spread can occasionally approach zero on liquid pairs, with a commission charged per lot.

Here is where the marketing gets loose. Very few retail brokers are purely one thing. Most operate a hybrid model, deciding order by order and client by client whether to internalise the flow or pass it out, a practice usually described as A-book and B-book routing. Profitable, consistent clients tend to get routed out; the rest are often internalised, because that flow is more valuable kept in house. An account labelled ECN may in practice be an STP feed with a markup, or internalised whenever the desk prefers. The label is a marketing decision as much as an engineering one.

The conflict of interest is real and worth naming plainly. If a broker is your counterparty, your loss is directly its gain. That does not make market makers fraudulent; a well-run dealing desk is a legitimate, heavily regulated business, and internalising flow is how you get instant fills on tiny sizes that no interbank venue would bother quoting. But the incentive exists, and it is why regulation, published execution policies and your own measurement of fill quality matter more than which of the three words appears on the account page.

How to trade it, step by step

1. Read the broker’s execution policy document. Regulated firms publish one. It states whether the firm deals on its own account, whether it acts as principal or agent, how it selects venues and how it handles conflicts. This document tells you far more than the account page, and almost nobody reads it.
2. Identify what you are being charged, not what you are being called. Take the same instrument on each account type the broker offers and add spread plus commission for the size you actually trade. Use the spread cost calculator to convert pips into money. A raw account plus commission and a wider all-in spread frequently land within a fraction of each other, and which one wins depends on your size.
3. Check whether depth of market is available. A genuine ECN feed can usually show you the book: the bids and offers sitting at each level. If a broker markets an ECN account but no depth is available on any platform it offers, treat the label sceptically and judge on measured execution instead.
4. Log your fills against your intended prices. Record intended entry price, actual fill and timestamp for at least a few hundred orders. Then separate the results into positive and negative slippage. Genuinely neutral routing produces slippage in both directions; a strong one-sided skew is a cost that never appears on any fee schedule. See slippage and requotes for how to do this properly.
5. Test behaviour under stress, not in quiet hours. Place small orders at the London open, at the New York open and around a scheduled data release. What you are watching for is whether spreads widen proportionately to the market, whether orders still fill, and whether you get requoted or rejected. Everything looks excellent at midday in a dead session.
6. Compare the same setup on two brokers simultaneously. Run demo or small live accounts side by side on the same instrument and place matched orders. This is the only comparison that controls for market conditions, and the differences it exposes are usually larger than any comparison table suggests.
7. Match the model to your actual holding time. If you scalp, fill quality and spread dominate everything and a raw feed with commission usually earns its keep. If you hold for days or weeks, swap rates will dwarf the spread difference, and a broker with better financing on a wider spread can easily be the cheaper choice overall.
8. Check the account minimums and sizing before you commit. Raw-spread accounts often carry higher minimum deposits or larger minimum lot sizes. If those force you into positions bigger than your risk plan allows, the cheaper spread is irrelevant, size from risk using the position size calculator and pick the account that lets you do it.

Size every one of those entries with the position size calculator and check the trade is worth taking with the risk/reward calculator before you commit.

The conditions it needs

You measure execution instead of trusting labels

The only reliable way to compare execution models is with your own fill data on your own instruments at the times you trade. A few hundred logged orders tells you more than every marketing page combined, and it is the one comparison a broker cannot influence.

The pricing model matches your trade size and frequency

Commission is a fixed cost per lot while spread scales with it, so the crossover point between a raw account and a standard account depends on your size. Work out where your typical trade sits relative to that crossover rather than assuming raw is cheaper.

The broker is regulated and its conflicts are disclosed

A market maker operating under a strong regulator with a published execution policy, best-execution obligations and monitored order handling is a very different proposition from one operating offshore with none of those constraints. Regulation is what makes the dealing-desk model tolerable.

Behaviour under stress is proportionate

Good execution is not defined by tight spreads in calm conditions. It is defined by whether the platform still works, whether fills still arrive and whether the widening you see around news is in line with what the broader market is doing rather than dramatically worse.

Your strategy is not fighting the model

High-frequency scalping on a wide all-in spread is a losing arrangement regardless of broker quality, and paying commission on every entry when you hold trades for three weeks is money wasted. Aligning the two removes a cost problem before it starts.

When it fails

• Assuming ECN means no conflict of interest. An ECN or STP broker earns from volume, which creates its own incentive: to encourage more trading, larger sizes and higher leverage. That is a different conflict from the market maker’s, not an absent one, and overtrading has ruined more accounts than dealing desks ever have.
• Believing the label describes the routing. Most retail brokers run hybrid A-book and B-book models and decide per order which route to use. An account marketed as ECN may be internalised much of the time. Only measured fill data and the published execution policy tell you anything reliable.
• Chasing zero spread and ignoring commission. A headline spread from 0.0 pips almost always sits on an account with a per-lot commission, and the from is doing heavy lifting. Add both together for your instrument and size before concluding anything about cost.
• Blaming the execution model for a losing strategy. Traders who lose consistently often conclude their broker is hunting their stops. Sometimes there is a genuine execution problem, and it will show up clearly in logged fill data. Far more often the stop was placed at an obvious level where the market was always going to look, which is a liquidity issue rather than a broker one.
• Ignoring what happens when liquidity disappears. In genuinely disorderly markets no execution model saves you. Raw feeds gap because there is nothing to fill against; dealing desks widen or stop quoting. Anyone relying on a stop being filled at exactly their price during a shock has misunderstood what a stop order is.
• Switching brokers on a small, emotional sample. Three bad fills after a losing week is not evidence. Traders who broker-hop on feel rather than on logged data pay the switching cost repeatedly and never accumulate enough data on any one venue to reach a real conclusion.

For different levels of experience

If you are brand new

Do not let this topic become a rabbit hole. As a beginner the execution model is far less important than being with a properly regulated broker whose withdrawals work. Plenty of profitable traders use market makers and plenty use ECN accounts.

The one thing worth understanding is the pricing difference. A standard account bundles the broker’s fee into a wider spread, so there is one number to think about. A raw or ECN account shows a very tight spread and charges a separate commission per lot. Neither is automatically cheaper, add them up for the trade you actually intend to place.

Be aware of the conflict of interest with a market maker, because it is real: if the broker holds the other side of your trade, your loss is its gain. Regulation is what keeps that in check, which is why the licence check comes first. See how to choose a forex broker.

If your results are inconsistent

This is the stage where traders start suspecting their broker. Sometimes they are right. Usually the fix is to replace suspicion with data: log intended price, fill price and timestamp for every order over a couple of months, then look at the distribution of slippage. If it is roughly symmetrical, your execution is fine and the problem is elsewhere. If it is heavily one-sided against you, you have a measurable case and a reason to move.

The other useful exercise is to work out your true cost per trade and put it into your expectancy calculation. Many inconsistent traders discover that a strategy that looks marginal on paper is actually fine before costs and negative after them, which is a cost problem with an obvious solution, not a strategy problem.

One more thing to check: whether your account type has changed. Brokers restructure account tiers, and a spread that was competitive when you opened the account may not be now. Re-price it annually against a live alternative.

If you are experienced

At size, the questions become concrete. Who are the liquidity providers, is last look applied and with what hold time, what is the rejection rate, how is the aggregated book constructed, and what is the internalisation policy for flow like yours. Ask directly; the answers you get, and the willingness to answer at all, are informative.

Measure rather than infer. Effective spread paid on filled orders, mark-out over the seconds after the fill, rejection rate by time of day, and slippage skew segmented by order type and by session. Mark-outs are particularly revealing: consistent adverse price movement immediately after your fills tells you something about how your flow is being handled that no policy document will.

Also watch out for tiering. Profitable clients frequently get routed differently from the rest, which means your execution can change materially without any announcement and without your account type changing. Track your fill statistics over time rather than benchmarking once at onboarding, and keep a second regulated venue live so you always have a comparison running.

Risk management for this strategy

The execution model changes the shape of your worst case more than your average case. On a raw feed, a shock produces gaps: your stop becomes a market order and fills wherever liquidity exists, which can be far from your level. With a dealing desk, the same event produces widening, requotes or rejections. Either way, a stop is an instruction to exit at the best available price, not a guarantee of a price, and any risk plan that assumes otherwise is understating its own tail.

Size accordingly. Work out what a gap of several times your stop distance would do to the account and confirm you can survive it, particularly on instruments that gap over weekends or around scheduled events. If the answer is uncomfortable, the position is too large regardless of how good the fills have been so far.

The other risk is the volume incentive. ECN and STP brokers profit from turnover, and the account structures around them (tight spreads, commission rebates at higher volumes, high leverage) all nudge in the same direction. Overtrading is the most expensive thing most traders do, and it is worth noticing when the fee structure is quietly encouraging it. Position sizing is the control that keeps that pressure from reaching your account.

Where Market Structure Pro fits

The execution model determines what your trades cost; it does nothing about how many of them you take. That second problem is the one Market Structure Pro is built for, and on a raw-spread account it is worth more than it looks, because commission is charged per trade whether the trade was worth taking or not.

MSP is spread-aware, reading the live spread on the chart rather than an advertised average. On a raw feed that spread moves constantly (near zero in the middle of London, several times that at a rollover or into a data release) and MSP folds it into the verdict and the confidence figure rather than pretending conditions are static. Combined with the session awareness and the dedicated ranging filter, whose entire job is to return NO TRADE in chop, it tends to remove exactly the trades that lose to costs rather than to the market.

It also gives you something to measure against. Because MSP is non-repainting and locks state on the closed bar, you can pair its verdict and A/B/C grade with your logged fill data and ask a sharper question: are the fills on A-grade setups behaving differently from the rest, and is a particular session or broker responsible? That is decision support, not execution. MSP does not place trades, is not a signal service, and guarantees nothing.

Frequently asked questions

What is the difference between an ECN and a market maker broker?

A market maker takes the other side of your trade itself and manages the resulting exposure internally, profiting from the spread and, indirectly, from client losses. An ECN broker passes orders into a shared electronic pool where other participants fill them, and charges a commission instead. STP sits between the two, routing orders to selected liquidity providers.

Is my broker trading against me?

If it is a market maker, it may literally be your counterparty, and its interests are opposed to yours on that position. That is legal and normal under a strong regulator, which imposes best-execution and conflict-management duties. Whether it is affecting your fills is a question you answer with logged data, not suspicion.

Is an ECN account always cheaper?

No. ECN accounts show a much tighter spread but add a per-lot commission, so the total cost depends on your trade size and the instrument. On small positions the commission can outweigh the spread saving. Add both components together for the size you actually trade before deciding.

What does A-book and B-book mean?

A-book means the broker passes your order out to a liquidity provider and earns from markup or commission. B-book means it keeps the trade internally and becomes your counterparty. Most retail brokers run both and route order by order, often sending consistently profitable clients to the A-book and internalising the rest.

Which execution model is best for scalping?

Scalping is the style most sensitive to spread and fill quality, so a raw or ECN-style feed with commission usually suits it better, provided fills are reliable and rejections are rare. What matters more than the label is measured execution during the hours you actually scalp, since a tight average spread in quiet periods proves very little.

Do ECN brokers have a conflict of interest?

Yes, just a different one. They earn from volume rather than from your losses, which creates an incentive to encourage more trading, larger positions and higher leverage. No model removes conflict entirely; the question is which conflict you are exposed to and whether regulation constrains it.

Why did my stop fill far worse than my stop price?

A stop order becomes a market order once triggered, so it fills at the best price available at that moment. In fast or gapping markets that can be well beyond your level, on any execution model. This is normal market mechanics rather than broker misconduct, and it is why position size should assume worse than your stop distance.

How can I tell if my broker is really an ECN?

Look for genuine depth of market on the platform, read the published execution policy to see whether the firm acts as principal or agent, and log your own fills to check whether slippage occurs in both directions. If depth is unavailable and the policy says the firm deals on its own account, treat the label as marketing.

Should I switch brokers if I get bad fills?

Only on evidence. Collect intended and actual fill prices across a few hundred orders and look at whether the slippage is genuinely one-sided against you. Switching on the back of a handful of bad fills after a losing week usually costs more than it saves, because you never accumulate enough data anywhere to know.

Related reading

• Slippage and requotes explained: How to measure fill quality properly instead of guessing.
• Understanding trading costs: Add spread, commission, swap and conversion together to compare account types.
• How to choose a forex broker: Where execution model fits into the wider broker decision.
• Spread cost calculator: Turn spreads and commissions into the money they actually cost.
• Broker comparison: Compare account types, pricing models and regulated entities.