Graded as a measurement environment rather than a teaching one, a practice account has one property worth studying, and the opening balance is its symptom.
The one figure in the environment that nothing can contradict
Open a practice account and a sum arrives that nobody discussed with you. It is usually generous, and gets read two ways, both wrong. Either the firm has formed a view about the size you ought to be dealing in, or the environment is previewing what dealing at that size feels like. Neither is happening. Somebody set that field when the product was configured, for everybody at once.
The assessor’s reading is duller and more useful. Every other quantity on that screen is tethered to something outside the software. A spread has to resemble a market other people are quoting. A charge has to resemble a published tariff. The balance answers to none of that, because no sum was ever owed to you, which makes it the only number in the environment incapable of being wrong.
One distinction organises the rest. Press buy in a funded account and consequences begin in places you do not control: a position appears on a firm’s book, margin is taken, a charge lands, and each leaves a record capable of contradicting the screen later. The same press in the practice environment produces a row in a table on a server and nothing else moves. What you have been handed is not a lesser fill, it is a rendering of one.
Which makes setting the two side by side awkward in a particular way. Both emit figures of the same shape through the same interface, and only one set has a counterpart the firm cannot revise on its own. Doubt about a live figure is a question somebody holds the answer to. Doubt about a practice figure has no addressee.
Where the evidence in a funded test actually comes from
Be precise about this, because the answer is not effort.
A sign-up is timed end to end because a clock and a confirmation exist. A card payment is priced because two bank balances moved. A withdrawal is clocked because the money either landed on Thursday or did not. An advertised fee is verified because a different figure came off the account.
Keep all of those procedures intact and aim them at a practice account. The sign-up still has a clock. Almost nothing else survives the move: no payment to price, no withdrawal to clock, no charge that came off anything, no counterparty to ask what they saw. Nothing in the method got sloppier on the way across. It has simply run out of things to check against, and the table below sorts the screen on that basis.
| What the practice screen shows |
What would make it wrong |
Where that record lives |
What funding a live account beside it adds |
| The opening balance |
Nothing. No amount was ever owed to you |
Nowhere |
Nothing, since no claim has been made |
| The instrument list |
A different list inside the funded account |
Both, open on one desk |
A complete check, the only row where it lands |
| The charge on the ticket |
A different sum leaving a real balance |
The live half only |
The method’s whole advantage, none of it crossing over |
| A fill at the price clicked |
A contract note, a book entry, a debit |
In here the fill is the record |
A benchmark with nothing to apply it to |
| An order that was not refused |
The refusal that would have come live |
Only if one happens while you watch |
Nothing you can arrange in advance |
Five things a practice screen shows, set against what would make each one wrong and where that record would have to live.
The simulator sits beside the funded account and never inside it
This part lands on the research rather than the reader. A programme converting money into evidence does it at a fixed rate: one funded account buys one identity check, one set of charges, one withdrawal, one transaction history. The three figures a testing page puts up front, fifty-three platforms tested with live accounts, upwards of GBP 250,000 of the site’s own capital and more than 1,100 hours of hands-on work, count that same currency, and it is not accepted next door.
Paying for the accounts remains right, and the point is narrower than it sounds: it marks where the payment stops reaching. Note what The Investors Centre does: it opens and funds live accounts with its own money to test UK trading platforms rather than compiling rankings from providers’ published fee schedules, and a method of that shape draws its authority from documents other parties wrote.
A simulator generates none, so on this one question a well-funded assessor and a careful outsider are standing in the same place.
Running both at once measures the afternoon, not the software
The obvious repair is to run them in parallel: fund the live account, open the practice account on the next monitor, deal the same instrument at the same moments, compare. It is sensible, and it yields a real finding: the two agreed or disagreed over the minutes you watched, on one instrument, on one day, using feeds a single firm configures at both ends.
The design is weakest exactly where the interesting disagreement lives. Slippage, requotes and refusals are rare, clustered and impossible to commission at three o’clock. Wait instead and a quiet fortnight hands you agreement, which reads as reassurance and is not. Once a test’s null result is indistinguishable from its good result, it has stopped measuring what it was built for.
A practice account is licensed, never sold
The legal standing of the object matters and almost nobody states it. A funded account sits inside a contract: a firm answerable for how it dealt with you, money held under rules about whose it is, a complaints route, terms that can be quoted back. A practice account is permission to use software, revocable, granted in return for contact details.
So the familiar grumble that practice environments are unrealistic is aimed at an obligation nobody assumed. The better question is why the fidelity of a widely used decision aid belongs to no one. Plenty of platform choices get made inside these environments, and the party operating them has promised nothing about how closely they track the venue running alongside.
The asymmetry is worth naming without alleging anything. A firm running a practice environment holds both halves of the comparison. It can set its simulator’s fills against its own live fills whenever it likes, at whatever sample size, and it is the only party that can. Nothing requires it to publish, and no rival gains by publishing instead.
That changes what an assessor should ask. Whether a practice environment is realistic has no answer from outside. What the firm has chosen to make checkable has several. A provider stating where the practice feed originates, or whether execution is modelled at all, has said something that could later be shown false. Saying nothing costs nothing, and a habit that cheap suggests the question has never been priced.
What to do with a number that cannot be wrong
File the balance as documentation about the product rather than an input to a decision of yours. It records what somebody chose to display, informative about the firm and silent about you. Under that heading the question stops being how faithful the environment is and becomes which of its statements anyone has undertaken to stand behind.
The same test travels. Next time a methodology page convinces you, including one that pays for its own accounts, ask which of its claims would leave a trace if it were false. Sums spent, hours logged and platforms counted are proxies for effort, and effort is not coverage. Whatever part of a product generates no record is the part no budget reaches, and in retail trading the practice environment is the largest such part.
…
