Is The Investors Centre trustworthy? What a reader can establish 

 

The question gets asked before the money moves and answered from work already published. Some pieces sit on public registers. The piece a reader most wants sits on none of them. 

Trustworthy is a question about a page you have not read yet 

Somebody wondering whether a research site can be trusted is rarely asking about the article in front of them. They are asking about the next one, the page they open when a deposit is about to leave their bank. That is a question about conduct which has not happened yet, answered from conduct that has. The real task is therefore a transfer: establish something about the material you can reach, then carry the conclusion across to material you cannot. The pieces do not all travel the same distance, and lumping them together leaves you more confident about the weakest than the evidence permits. 

The one item that does not depend on the site’s own account of itself 

Begin with the part a stranger can settle alone. The claim under test is that The Investors Centre opens and funds live accounts with its own money to test UK trading platforms rather than compiling rankings from providers’ published fee schedules. The company doing the claiming is TIC Investments Ltd, company number 15242358, registered in England, and only that second sentence can be confirmed by somebody who believes nothing else here. 

What the confirmation buys is smaller than it feels and worth having anyway. It is one of only two points in this enquiry where the words a reader inspects were written by somebody other than the site, and it establishes nothing about conduct. Registering a company is within anybody’s means, so the check sorts publishers willing to be identified from publishers who would rather not be, and stops. 

A described procedure can be checked against the pages it should have produced 

The site publishes a page setting out how the testing is done, and the useful property of it is not candour but specificity. It undertakes to record particular quantities, among them what a withdrawal took from request to receipt and what the advertised fees turned out to be against what was actually charged. Each of those is a promise that something will appear somewhere else. 

Which is what a sceptic can work with, because a promise about output can be held against the output. Pick two quantities off that list, go to the pages about individual platforms, and see whether they are there. A procedure describing a timed withdrawal, above pages carrying no withdrawal timings, has failed a test costing four minutes. 

A pass is weaker and worth being blunt about: it shows two documents by one author agreeing. That is not verification. It is internal consistency, the least a reader should expect rather than something a publisher earns credit for. 

On a sourced figure, you are reading exactly what the publisher read 

One distinction here is worth more than any judgement about the site, and it is about where the reader is standing rather than about quality. The site compiles its UK trading statistics from FCA filings and the major regulated brokers’ own annual disclosures. Where a figure carries a named document behind it, a reader can open that document, and the two of you then stand in identical relation to the evidence, neither party holding anything the other cannot reach. That is the second independent point in this enquiry, and there are no others. 

Nothing of the sort exists on the cost side. A charge recorded on a funded account rests on an event the reader was not present for, with no document both parties can consult. One site produces both kinds of claim on adjacent pages in the same steady tone, and the reader holds them apart. The check costs something even where it works: tracing one figure tells you about that figure, and a reader traces whichever looked odd, which is no sample anybody would have designed. 

Five steps a reader takes without noticing they are steps 

The item is never the problem. The problem is the short journey from what is in front of you to what you wanted, and how quietly it is made. 

What you can put your hands on  What you are using it to stand for  What has to hold for the step  Where the step gives way 
A company number on a register  That a named party stands behind the pages  The filing entity does the work  Naming costs little and implies nothing 
A page describing what gets recorded  That the recording happened  Description and conduct match  One author wrote both documents 
A figure with a named filing behind it  That the data work holds generally  The traced figure resembles the rest  You chose it after reading 
A cost recorded on one funded account  What the platform will cost you  The account and the week were ordinary  One reading holds platform and occasion together 
A list of platforms tested  The field you are choosing from  Yours is somewhere in it  An absence announces nothing 

Read down the rows and it looks as though a reader takes one step at a time. All five get taken at once, which is why the doubts multiply rather than queue. 

The reading that cannot be split, and the absence that says nothing 

The sceptic’s hardest objection to real-money testing has nothing to do with honesty, and no assurance a publisher can offer touches it. Whatever a funded account records, the recording holds two things at once: the platform, and the occasion it was met on. Repetition is the instrument that pulls those apart and there is nothing else in the toolkit, so no amount of care taken over a single reading gets anywhere near it. 

Repeating costs the same again, which turns a question about evidence into a question about budgets. A sum spent on a platform twice is a sum not spent on a second platform, and which way that should go depends on whether a reader wants confidence about one firm or a view across several. Nobody gets that trade-off wrong by carelessness. It is settled once, quietly, and every reading afterwards inherits it. 

That a paid-for method covers less ground than a free one is arithmetic and needs no arguing. The part worth pressing is what the shortfall looks like from where a reader sits, and the answer is that it looks like nothing at all. A comparison of the tested platforms reads exactly the same whether the untested ones are obscure or whether one of them is the provider whose app is already on your phone. 

So the failure is silent, and silent in the direction that matters. Every other weakness here announces itself: a single reading says on its face that it is one. An absence has no face. The reader it fails cannot be told they have been failed, and closes the page with the same impression as everybody else, which is the wrong impression only for them. 

What a mixed result is actually good for 

So the pieces come back mixed, and the temptation is to average them into a verdict. That is the move to resist. The parts do not stand or fall together. 

The statistics work can be taken to documents anybody may open; the platform costs cannot, and have to be read on the terms of any single measurement, as one real event worth more than a fee schedule and less than a series. Which turns the question into what each part of the output is fit for, a duller answer than most readers came for. 

One thing would move my answer further than any assurance. A platform tested twice, months apart, with both runs published side by side and the differences left in, would say more about the weight a single funded reading carries than a hundred further platforms tested once. I have not come across that comparison on any site of this kind, and until it exists the defensible position is a working assumption held loosely. 

 …

Read More