We opened an account at IC Markets and sampled it every second

How IC Markets gets to $910 a year

The headline figure is one number, and one number is not checkable. Here it is in its parts, for 10 lots a month of EUR/USD on the Raw Spread · MT5 account we opened:

One pip on one standard lot of EUR/USD is about $10; that conversion is the only constant in the sum, and it is the market's, not ours.

On this account the commission is the larger half of the bill: $840 of $910, or 92%. That is the thing a spread table cannot tell you. IC Markets quotes a raw spread of 0.06 pips, which looks almost free, and then charges per lot for the privilege. Whether that is cheaper than a wider spread with no commission depends entirely on how much you trade, which is why the profile is stated rather than assumed.

One thing the bill above deliberately leaves out is the rollover. Between 21:00 and 24:00 UTC the spread on this account averaged 0.84 pips against 0.06 in the session — 15 times wider — and folding those three hours into a daily average would have told you about a market nobody trades in rather than about the price you pay. It is measured and printed as its own column. If your positions cross midnight UTC, that column matters more to you than the headline.

Nothing in this table is a quote. It is what our own account would have paid at a stated volume over the window we measured, and volume is the lever that moves it most. Both components scale with volume, and because the brokers here differ in which of the two dominates, the ordering is a function of how much you trade.

The rig behind IC Markets's numbers

The figures come from a single Raw Spread · MT5 account at IC Markets — retail terms, public sign-up, nothing arranged on the phone. The terminal runs on a dedicated machine that stays awake around the clock. That sounds trivial and is not: the out-of-the-box power settings would have silently deleted the overnight window from every series.

Ticks land in the store unmodified, and the one transformation applied to them — the shift from broker-server time to UTC — is measured per session rather than assumed. The guard that holds whatever the offset says is freshness: a tick has to be seconds old once converted, or it does not enter the store. Without it a stale weekend quote reads as a live one.

No execution figure survived from this account. The reason is printed on the axis itself rather than hidden: an execution figure from a handful of orders is noise with a decimal point on it, and we would rather show nothing than that.

2026-10-12T19:08:17Z to 2026-10-13T17:13:27Z, continuously, with coverage counted against calendar market hours rather than against the hours we happen to have seen.

How a price series earns its place here

The danger with a demo account is not a wrong number but a fictional one. Averages, hourly breakdowns and rankings built on such a feed measure the simulator rather than the broker, and every figure derived from it inherits the fiction while looking exactly like a measurement. So before a single number from an account is published, its series is tested for one property: continuity.

A floating spread moves in steps of one unit, so it is seen at the values either side of its usual one. If a series usually reads one unit wide it must sometimes read zero and sometimes two; if it never visits its neighbours it is being set rather than formed. The test is run on the liquid session only, because an off-session widening schedule can hand a fixed daytime price exactly the neighbours it lacks — that is how one feed passed a day-wide version of this audit and failed the session-only one.

The test has teeth, and the refusals are published. 8 series across 2 accounts failed it for this window and are not represented anywhere on this site — not as a row, not as a footnote figure, not as a “from” number. One of them belongs to the broker that pays for the site.

Execution quality is the next thing this rig will measure on these same accounts; until there are numbers, there is no column, and that is deliberate.

Spread by hour on the Raw Spread · MT5 account

Every hour of the day, as measured on this account: the mean of one-second samples falling in that hour across the whole window. The widest hour is 21:00 UTC at 2.391 pips, the tightest is 17:00 UTC at 0.050. 16 hours of the day were recorded on both this account and the others it is compared with; the rest are not shown, because an hour measured on one account only is not comparable with anything.

The rollover window, 21:00-24:00 UTC, averages 0.844 pips against 0.058 through the London and New York hours — about 15 times wider. That ratio, not the daily average, is what decides whether an overnight strategy is viable on this account, and it is the number a marketing page never carries.

Hourly mean of 1 s bid/ask samples · 12.10 19:08–13.10 17:13 UTC · our own Raw Spread · MT5 account · how the hours compare across brokers

What IC Markets's licence gives a client

The protections attached to IC Markets in this table are the ones that come with 362/18, held by IC Markets (EU) Ltd. They are properties of that entity, and they are worth reading before any spread figure on this page.

Read together, those four lines describe a fairly specific deal. Your losses on a CFD account cannot exceed the money in it, because negative balance protection caps aggregate liability at the account balance. Positions are closed out when equity falls to half the required initial margin, which is a floor under how far a bad position can run before somebody intervenes. Leverage on major currency pairs is capped at 30:1 — the same broker group may legally offer several hundred to one through a different entity, and that is one of the clearest practical differences between the licences a single brand can hold. And if the firm itself fails owing you money, the compensation scheme pays up to €20,000, which is a ceiling worth knowing against the size of a deposit rather than after it.

None of that is a judgement about IC Markets. It is the floor the licence puts under any firm holding it, and the reason this site records the entity rather than the brand: the measurements above would be identical under a different entity, and the protections would not.

Stated by the regulator · read on 2026-10-11 · we quote the rule, not a summary of it, because a summary is where this kind of fact usually goes wrong

One caution about reading protection as cost. The regulated European entity is often not the cheapest venue a group operates: the same brand's offshore entity may quote tighter and lever higher. Our figure of $910 a year is for the entity above, on the account type above, and it is not transferable to a differently licensed arm of the same company.

4 axes, 4 placings for IC Markets

IC Markets is in the better half of the table on 1 of the 4 axes it has figures for — last of 3 on annual cost; last of 3 on avg spread, 07:00–20:00 utc; best of 3 on spread 21:00–24:00 utc. Each placing is against the other accounts in the same source class, on the same instrument and window - not against a broker's claim, and not rolled into a composite score. A composite score is exactly where a site paid by one of these brokers would hide its thumb, so there is none.

Lower is better on every axis · 12.10 19:08–13.10 17:13 UTC · ranking happens inside one source class only

What this column cannot tell you is which axis matters to you. That depends on volume, holding period and the hours you trade, and we are the last people who should be weighting it on your behalf.

Questions a reader usually asks next

Each answer is derived from the figures on this page. None of them is a general statement about brokers.

Where to point your scepticism

We are paid by one of the brokers in these tables. That is a reason to check the work rather than to trust it, and these are the places where checking is cheapest.

• Recompute the billthe components and the profile are printed above; the arithmetic is deliberately simple enough to redo on paper
• Ask for the rowsthe data page says what we hand over, including the flagged readings we excluded
• Compare us with anybody else measuring the same thinga figure that only exists on one site is a figure nobody has checked, ours included
• Check the window, then the sort orderevery figure carries the window it was taken in; the cost table is sorted by cost, and our sponsor sits wherever that puts it
• Check that the losses are printedif the sponsor's losses ever stop appearing on this site, stop believing the rest of it

If you find something wrong, tell us: [email protected]. We publish the correction and keep the wrong rows, flagged and excluded, because a record of our own faults is worth more than a clean-looking database. Who may alter a published figure, and what happens when a broker disputes one, is written out in the editorial policy.