Performance claims on social media

What a screenshot actually is 

A screenshot is an image of a screen at a moment, and that is the whole of its evidential content. It records what a display showed, not what happened in an account, and those two things come apart in several entirely ordinary ways. Figures rendered in a browser can be altered in the browser's own developer tools in seconds, with no specialist skill and no trace left in the image. A simulated account shows the same interface, the same instruments and the same profit column as a funded one, which is the point of a simulated account rather than a defect in it. An image of an open position says nothing about the positions closed the same morning, and an image of a closed one says nothing about the money paid in to fund it.

The second problem is the crop. A screen is a window onto a record, and what sits outside the window is not contradicted by the image, it is simply absent: the other open positions, the transaction history, the transfers in and out, the balance a month earlier. An image also cannot be questioned. There is nobody to ask what the rest of the record contained, no way to establish that the account exists, that the person posting it holds it, or that the entries on it were placed at the times displayed.

A statement is a different class of object. It is issued by the firm that holds the account, in the name of the account holder, covering a stated period, and it reconciles: opening balance, every transaction, every adjustment, closing balance. It exists in the firm's records as well as the holder's, it can be produced to a regulator, and a discrepancy in it is something the firm is answerable for. None of that is true of a picture. That is the distinction the word verification carries. The question is not whether a figure looks plausible. It is whether an independent party holding their own copy of the record can be asked to confirm it.

The denominator nobody posts 

Now assume the harder case. Assume a posted result is genuine, belongs to the account it came from, and has not been touched. It still carries very little information, because the number that would make it informative is the one that never appears: how many attempts stand behind it, and how many other people made the same attempt and are not posting. A published outcome is drawn out of a pool of outcomes by the person publishing it, and that selection is not random. The good ones are posted. The bad ones are deleted, or were never posted, or belong to people who went quiet.

This is not primarily a claim about dishonesty. Selective reporting happens with no intent to deceive at all, because posting a good outcome is pleasant and posting a bad one is not, and because an account that produces nothing worth posting stops posting and disappears from view. The effect is identical either way. What remains visible is the surviving tail of an invisible pool, and a tail produced entirely by chance is indistinguishable, item by item, from a tail produced by ability. The arithmetic below is worth reading slowly, because it is the whole lesson in one table.

Nothing in that table involves judgement, method or a market. It is a coin. And at the end of it there is a sequence with an unbroken run behind it, which, presented on its own without the thousand and twenty three that broke, looks exactly like evidence of a method. The failures are not concealed by anybody. They are merely not interesting enough to publish, and nobody counts what was never put in front of them.

The same arithmetic, run deliberately 

The pattern has a deliberate version, older than the internet and worth describing because it turns the accident above into a technique. Rather than one sequence surviving by chance, a list of recipients is split in half and opposite predictions are sent to each half. Whatever the market then does, half the list received a correct prediction. The half that received the wrong one is dropped, and the remainder is split again.

The people left at the end have received an unbroken run of correct calls, every one of them checkable against the market and none of them fabricated. They are nonetheless looking at the output of a splitting rule rather than at a forecaster, and no amount of checking the individual calls can reveal it, because the individual calls are all true. The only thing that would reveal it is the denominator, and the denominator was discarded at every round.

What can be changed after the event 

Order is the other thing a social platform records badly. A market call is only a call if it existed before the move it describes, and almost nothing on a social platform establishes that reliably to an outside reader.

• A post can be edited after publication, and on many platforms the edit history is not visible to a reader.
• A post can be deleted, which removes it from anyone else's count while leaving the poster's own sense of their record intact.
• A private message displayed as a screenshot can be composed after the move it appears to anticipate.
• Opposite calls can be posted in different places, and only the one that aged well referred to afterwards.
• A timestamp inside an image is a property of the image, and an image is not a timestamp.

None of this requires sophistication, and none of it leaves a mark. It is worth separating from fraud, because most of it is not fraud: a person tidying their own feed is doing something unremarkable, and the record that results is misleading anyway. The mechanism does not need bad intent to work.

What a licensed communication has to carry 

A communication from a licensed firm is a regulated document, and it stays one whether it appears in a prospectus or in a post. Obligations differ between jurisdictions, but the same furniture recurs across them. The firm is identified by its legal entity rather than by a brand or a handle. Its regulator is named. The material states plainly that trading carries a risk of loss. Marketing material is identifiable as marketing rather than as neutral commentary, and educational material is labelled as education rather than as advice. Underneath all of it sits one standard that regulators express in almost identical words, that a communication be fair, clear and not misleading, and that standard governs what is left out as firmly as what is said.

Where a rulebook permits past performance to be shown at all, it is heavily conditioned rather than free. A stated period, a stated source, the same basis applied across the whole period, the effect of costs, and a statement that past performance does not indicate future results are the usual conditions. They exist precisely because the arithmetic in the previous sections is well understood by the people who write the rules. A figure carrying none of that is not a weaker version of a compliant performance figure. It is a different object, which a licensed firm would not be permitted to publish in that form.

That yields a structural test which does not depend on assessing the content at all. A communication carrying none of the required furniture was either not produced by a licensed firm, or was produced by one that is not treating it as a communication it is answerable for. A risk statement is part of that furniture, and the one this site carries reads as follows.

Promotion is itself a regulated activity 

Advising on financial products, arranging deals in them and promoting them are, in most regimes, activities that require a licence or an exemption, and the requirement attaches to the activity rather than to the job description of whoever performs it. General explanation of how a market works is not usually a regulated activity. A specific recommendation aimed at an audience, or the promotion of a particular firm's services, moves toward one. Regulators describe that boundary as fact-specific rather than bright, which is a candid way of saying that people who cross it frequently believe they have not.

Payment is the least visible part and the most consequential. Marketing arrangements are ordinary across the industry: a person who introduces clients to a firm is commonly paid for it, sometimes per account opened and sometimes as a share of the costs those clients go on to generate. Such an arrangement is not fraudulent in itself and is not hidden from regulators, who generally require it to be disclosed to the audience as well. The disclosure is the part that matters, because payment tied to account openings or to trading volume creates an interest in more accounts and more trading that exists whether or not anything being shown is real. Without knowing how a person posting is paid, an audience is missing the single fact that best explains what is being posted.

The position on the unlicensed version is not softened by the material being free of charge. Somebody who is neither licensed nor exempt and who promotes trading in financial products sits outside the regime, and that has a practical consequence beyond the legal one. The conduct rules, the supervisory relationship and the complaint route described earlier in this module attach to licensed firms. They do not attach to an account posting results, and there is no version of them that does.

Where practitioners disagree 

Independently verified records are the serious counterargument, and the disagreement about them is real. One position holds that the problem is solved by third party verification: an independent service reads an account's history directly from the platform, the holder cannot edit what it reads, and what comes out is a genuine record rather than a picture. That is true as far as it goes. The other position is precise about what verification establishes, which is that those entries occurred in that account, and about what it does not, which is that the same approach produces anything in anyone else's hands, that the account was not chosen from many, or that the period shown resembles the period ahead. Verification defeats fabrication. It does not defeat selection, because whoever decides which verified account to publicise is doing the selecting.

A second disagreement runs deeper, about whether unlicensed commentary is a harm at all. One view is that explaining how markets work is a public good regardless of who provides it, and that a licence has never been a proxy for whether an explanation is correct. The other is that an audience cannot separate the explanation from the recommendation embedded in it, and that the economics of attention reward the recommendation. Both positions are held by serious people. Neither of them has anything to say about a figure posted with no denominator, which is the narrower thing this lesson is about.