Does SMC actually work? What it can and cannot claim?

Smart Money Concepts describes real things on a chart. What it cannot do is prove the story it tells about them. Separating those two — the description and the explanation — is the difference between using SMC as a tool and believing it as a doctrine.

We have written thirteen articles explaining SMC, so this one is not a takedown. It is the piece we owe anyone who read the other thirteen: what the framework can honestly claim, what it cannot, and how much of it is new.

What SMC gets right?

The vocabulary points at things that genuinely exist on a chart, and you can verify every one of them yourself:

• A fair value gap is a measurable imbalance between wicks. It is there or it is not.
• A break of structure is a definition applied to prior highs and lows. Mechanical, and reproducible.
• Price does cluster around prior supply and demand areas, and stops do sit above old highs, which is why liquidity pools there.

None of that is in dispute. As a language for describing where price has been and where orders are likely resting, SMC is precise — arguably more precise than the vague "support zone" it replaced. That precision is a real contribution.

What is not proven?

The claim that does not survive scrutiny is the causal one: that a bank left a pending institutional order inside that specific candle, and that price returned to fill it.

There is no published evidence for this. Not "the evidence is disputed" — there is none. And the reason is structural: institutional order flow in forex is not observable. Forex is decentralised and over-the-counter; there is no central tape, no consolidated volume, and no public record of who traded what at which price. A retail chart physically cannot show you what an institution did.

So the honest position is not "SMC is false". It is that the institutional story is unfalsifiable — it cannot be proven or disproven with the data any of us has. And an unfalsifiable explanation is a poor thing to build conviction on, because it explains every outcome equally well. Price respected the order block? The institutions defended it. Price sliced through it? It was not a "valid" order block, or it was an inducement. A model that cannot be wrong cannot teach you anything either.

How much of it is genuinely new?

Less than the marketing implies. Most of the SMC lexicon has a direct ancestor in work that predates it by decades:

• Order block — a supply/demand zone, itself a renamed support or resistance level.
• Liquidity sweep / stop hunt — Wyckoff's spring and upthrust, described in the 1930s.
• Smart money — Wyckoff's "composite operator", same idea, same century-old reasoning.
• Accumulation and distribution — Wyckoff, using those exact words.
• Break of structure — a higher high or a lower low.

Richard Wyckoff was writing about a large, informed operator accumulating stock before a markup in the 1930s. That does not make SMC worthless — renaming something can genuinely make it easier to teach, and SMC's rules are often tighter than the loose price action they replaced. But "institutional secret" is a marketing claim, not a historical one.

Why is SMC so hard to test?

This is the part that matters most for anyone deciding whether to trust their money to it, and it is a methodological problem rather than an ideological one.

SMC is discretionary. Which high counts as structure, which candle is the order block, which gap is worth taking — these are judgement calls, and two competent traders will mark the same chart differently. That makes a clean backtest nearly impossible: you cannot mechanically encode a rule that depends on interpretation, and if you loosen the rules until you can, you are no longer testing SMC.

It also makes the framework unusually vulnerable to hindsight. On a chart of the past, the order block that worked is obvious and the four that failed are invisible — you simply do not mark them. That is the ordinary mechanism of curve-fitting and confirmation bias, and no amount of screen time protects you from it. It is why "look how clean this setup was" is worth nothing without a forward-tested record.

So how should you use it?

The framework is not the problem. The claim attached to it is. Three adjustments make SMC honest:

1. Use it descriptively, not predictively. BOS and CHoCH confirm what already happened. That is genuinely useful for structuring an entry — it is not a forecast.

2. Drop the institutional narrative. You lose nothing operationally. A level either holds or it does not, and your P&L is identical whether or not a bank was involved. The story adds conviction without adding information, and misplaced conviction is expensive.

3. Demand the same evidence you would from any other method. A forward-tested sample, a known win rate, metrics you actually track. If a strategy only looks good in annotated screenshots of the past, it has not been tested.

The takeaway

SMC is a decent descriptive language wrapped in an unprovable origin story. Used as vocabulary for reading structure and liquidity, it is fine — and our other articles explain it on exactly those terms. Used as evidence that you are seeing what institutions do, it is a belief, and it should be held with the confidence a belief deserves rather than the confidence a fact would.

Anyone selling you certainty about institutional order flow is selling you something they cannot have.