Tickmill, measured rather than reviewed: cost, spread by hour, rollover

How Tickmill gets to $722 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 · 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: $720 of $722, or 100%. That is the thing a spread table cannot tell you. Tickmill quotes a raw spread of 0.00 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 1.12 pips against 0.00 in the session — 562 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. Whoever holds through the rollover pays that number rather than the headline one.

These are not quoted prices. They are what one real account would have paid over a stated window at a stated volume - and volume moves the total more than any difference between brokers. At a tenth of this activity both lines shrink to a tenth, and since one broker is cheaper on spread and another on commission, the ordering between them can invert.

Spread by hour on the Raw · 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 2.967 pips, the tightest is 02:00 UTC at 0.000. 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 1.123 pips against 0.002 through the London and New York hours — about 562 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 · MT5 account · how the hours compare across brokers

What we ran, and for how long

What sits behind these figures is one retail Raw · MT5 account at Tickmill, obtained the way a reader would obtain one. Its platform runs continuously on hardware we control, with sleep, hibernation and battery throttling all disabled — the defaults would have removed the thin-liquidity hours, which are the ones worth measuring.

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. Anything that is not seconds old after conversion is refused. That rule exists because a stale quote's age lands on a round hour roughly once an hour, so a whole-hour sanity check passes it cheerfully.

This account has no publishable execution sample. 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.

Collection ran from 2026-10-12T19:08:17Z to 2026-10-13T17:13:27Z without interruption, and what counts as coverage here is the share of market hours in the window holding real samples. The looser definition — hours of the day ever seen — once let a dead collector report itself healthy for three and a half days.

What we checked before publishing anything from Tickmill

A demo server can quote a price that looks entirely plausible and never moves. 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 real spread passes through its neighbours. 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.

Here is where a cruder test would have thrown away the truth. Tickmill quoted bid equal to ask in 98% of the samples we took in the liquid session: on a commission account the spread genuinely sits on the floor of the price grid, and the broker earns from the per-lot charge instead. An audit that refused series for being “too constant” would have condemned that as a units error — it is the first version of ours that did, and the test had to be rewritten around continuity rather than variety.

What the test refused is as public as what it passed. 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.

What a broker does with an actual order is measurable too, on the same accounts, and it will be published the same way: as figures with a window, or not at all.

What Tickmill's licence gives a client

The protections attached to Tickmill in this table are the ones that come with 278/15, held by Tickmill Europe 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 Tickmill. 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 $722 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 Tickmill

Tickmill is in the better half of the table on 3 of the 4 axes it has figures for — best of 3 on annual cost; best of 3 on avg spread, 07:00–20:00 utc; last of 3 on spread 21:00–24:00 utc. Placings are computed within one source class only, on identical instrument and window, and deliberately left unaggregated. 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

The spread of placings matters more than any single one: an account that is first on the bill and last at the rollover is a different proposition from one that is middling on both, and a single number cannot carry that difference.

Questions this page should answer

Answers come from this account's own numbers, and where a question needs a figure this rig has not produced, the answer says what producing it would take.

How to check what is on this page

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.

• Check that the losses are printedif the sponsor's losses ever stop appearing on this site, stop believing the rest of it
• Check the source class on each numberour measurement, third party, or stated by the broker — and no ranking crosses those classes
• 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

If you find something wrong, tell us: [email protected]. Corrections are published, and the bad rows stay in the store flagged — a tick cannot be collected twice, so deleting one destroys the evidence. Who may alter a published figure, and what happens when a broker disputes one, is written out in the editorial policy.