vs

AvaTrade vs FxPro

One instrument and one window, measured on accounts we hold ourselves. Green marks the better number on that row, not the better broker.

EUR/USD · 27.08–25.09 · profile 10 lots a month

Axis	AvaTrade	FxPro
Annual cost	$1 690🔬 our measurementderived · 27.08–25.09	$1 554🔬 our measurementderived · 27.08–25.09
Avg spread, 24 h	1.35 pips🔬 our measurement1 s bid/ask · 27.08–25.09	0.55 pips🔬 our measurement1 s bid/ask · 27.08–25.09
Spread 21:00–24:00 UTC	2.60 pips🔬 our measurement1 s bid/ask · 27.08–25.09	2.02 pips🔬 our measurement1 s bid/ask · 27.08–25.09
Commission per lot	$0.00📄 stated by the brokerfee schedule · checked 25.09.26	$7.00📄 stated by the brokerfee schedule · checked 25.09.26
Swap per year	$70🔬 our measurementMT5 swap log · 27.08–25.09	$49🔬 our measurementMT5 swap log · 27.08–25.09
Rejected orders	0.8%🔬 our measurement4 787 orders · 27.08–25.09	0.4%🔬 our measurement4 823 orders · 27.08–25.09
Slippage on releases	1.3 pips🔬 our measurement207 release orders · 27.08–25.09	0.8 pips🔬 our measurement214 release orders · 27.08–25.09
Withdrawal fee	$0📄 stated by the brokerfee schedule · checked 25.09.26	$0📄 stated by the brokerfee schedule · checked 25.09.26
Withdrawal time	no datanot measured in this release	no datanot measured in this release
Support response	no datanot measured · see methodology	no datanot measured · see methodology

AvaTrade is measured here and named in full. It is not linked — we have no commercial relationship with it, and we are not going to pretend otherwise by sending you there.

Putting the two columns into dollars

The table above is colour-coded by axis, which makes it easy to read and easy to misread: eight green cells on one side do not mean that side is cheaper or better, because the axes are not worth the same amount of money.

On annual cost of ownership, FxPro is $136 a year cheaper at 10 lots a month — $1 690 at AvaTrade against $1 554 at FxPro. Inside those totals, AvaTrade pays $1 620 in spread and $0 in commission, while FxPro pays $665 and $840. Swap adds $70 and $49 respectively, and that line moves with how long positions stay open rather than with how often they are opened.

On execution the order reverses. FxPro rejected 0.4% of our orders against AvaTrade's 0.8%, and slipped 0.8 pips on release-timed orders against 1.3. Neither figure can be looked up anywhere: producing them means sending thousands of real orders and keeping the broker's own status for each one.

Converting that into money depends on your strategy. A rejected order in a quiet hour costs a retry; a rejected order in the seconds after a rate decision, when the price has already moved, can cost more than a year of the spread difference above. The difference in slippage between these two accounts — 0.5 pips on a release order, roughly $5 per lot — is charged every time one is placed into that window.

Neither account is “better”. A high-volume day trader who is flat overnight should weight the commission and the rollover spread; somebody holding positions through news should weight slippage and rejects, where the cheaper account here is not necessarily the stronger one. Both columns are printed in the same type size because the weighting is yours to apply.

Each axis, and who is ahead on it

Eight axes, each with what it is for and which account is ahead on it. The list is deliberately flat — no weighting, no composite score — because the weights depend on how you trade, and a site paid by one of the two names here has no business choosing them.

Who the two of them are

The figures above belong to an account, the protections below to a company. Those are different objects, and the difference is the most consequential thing on a page like this: two clients of the same brand, onboarded through different entities, can have identical platforms, identical spreads and materially different protection if the firm fails.

For a confirmed entity, the broker's page quotes the regulator on four things: the compensation ceiling, the leverage cap, when positions are closed out, and whether a loss can exceed the deposit. An unconfirmed entity is printed as unconfirmed, and the measurements stand on their own either way.

How this pair was measured

The comparison is only worth reading because both sides went through the same pipeline at the same time. Figures taken weeks apart, or on different pairs, can disagree by more than the brokers do.

Here it is EUR/USD on both sides, 2026-08-27 to 2026-09-25 continuous, on the Standard · MT5 account at AvaTrade against the Raw+ · MT5 account at FxPro — both opened through the ordinary retail sign-up, both running on the broker's own MetaTrader 5 build on hardware that does not sleep.

The execution samples behind this pairing are 4 787 orders at AvaTrade and 4 823 at FxPro of which 207 and 214 respectively were timed into the minutes around scheduled releases. A rejection rate without an order count beside it is not a measurement which is why both are printed.

It is one instrument: EUR/USD is the most liquid pair in the market, and a broker that looks tight here can be much wider on a cross or a metal, where its own margin is less exposed. And it is one window: a month of measurement is a month, not a permanent property of either firm. That is why the window is printed at every figure instead of being mentioned once in a methodology nobody opens.

The same comparison at six volumes

The bill on either account is per-lot cost times volume plus a yearly swap, so a comparison at one volume answers one point of a line. Our headline profile — 10 lots a month — is one point on it, and a reader who trades a tenth or ten times that is reading the wrong point. Here is the same arithmetic at six volumes, with our profile marked.

Spread × $10 per pip per lot, plus commission per lot, plus a year of swap · 27.08–25.09 · the swap line does not scale with volume, which is what makes the lines cross at all

There is no crossover inside any realistic volume: FxPro is cheaper per lot traded ($12.54 against $13.50) and the difference in annual swap, $21, is not enough to reverse it. The ranking between these two is stable across volumes, which is unusual in this table.

Read it with two caveats: the spread is the one we measured on retail size, and spreads do not stay constant as size grows. It also assumes your hours look like the window average. Anything trading through the roll pays more than this table shows, on both accounts.

What the thin hours cost, in money

A bill computed from an all-hours average understates the cost for anybody whose orders land in the rollover window. AvaTrade goes from 1.35 to 2.60 pips in that window; FxPro from 0.55 to 2.02. Below is the same annual sum with a share of the volume moved into those hours.

Spread blended between the window mean and the rollover mean in the stated proportion · commission and swap unchanged · 10 lots a month

At a fifth of the volume in the thin hours, the penalty is $300 a year on AvaTrade and $351 on FxPro — FxPro carries the heavier one. No broker publishes this number, and it can be larger than the whole difference between the two accounts on cost.

This is a blend, not a simulation: your remaining volume is assumed to look like our average, and a single bad night around a release can cost more than the whole annual penalty above.

Both of them against the rest of the table

This pairing is a slice of a larger table, and a slice can hide the fact that both of its members are beaten elsewhere. So here are the same two accounts placed against every account we measure, axis by axis. AvaTrade leads the whole table on commission per lot; FxPro leads the whole table on rejected orders.

Placings inside one source class · 27.08–25.09 · the full ranking is on the brokers page

Reading the table in full is the better route for anybody choosing from scratch; a pairing is for a reader already weighing these two.

What this pairing does not tell you

Nothing here recommends either broker. It is a measurement of two accounts in one window, written by a desk with a commercial interest in exactly one of them — FxPro pays us, AvaTrade does not. Knowing that is part of reading it: the method is written out at length because the incentive runs one way and the only defence is making the figures checkable.

Plenty that matters is missing: how fast money comes out, what support does at three in the morning, how the platform behaves in a real panic, and how either firm handles a disputed fill. Some of that we have not measured; some of it cannot be measured from a retail account at all. “Which broker is better” is a much larger claim than anything on this page, and it is not one a measurement can settle.

Questions about this pairing