Home › Learn › Market Structure Basics on Gold
Market Structure Basics on Gold: Break of Structure, Order Blocks and the ICT Vocabulary
Break of structure, order blocks, fair value gaps, liquidity sweeps. Some of this vocabulary describes something real about how price moves. Some of it is a new name for a very old idea. And some of it cannot be wrong, which is the most dangerous category. Here is how to tell them apart on gold.
Start with the only thing that is objective
Everything in market structure rests on swing points, so define them before anything else.
A swing high is a bar whose high is higher than the highs of some number of bars either side. A swing low is the mirror. The number is your choice — two bars either side, three, five — but you have to make that choice and then keep it.
This sounds trivial and it is where most structure analysis quietly falls apart. If your definition of a swing changes depending on what you want the chart to say, then every conclusion downstream is unfalsifiable. Two traders looking at the same gold chart will identify different structure, and neither can be shown to be wrong, because neither wrote down the rule.
So: pick a swing definition, write it down, apply it mechanically. If a level only counts as structure when it supports your bias, you are not reading structure.
Break of structure and change of character
With swings defined, the rest of the trend vocabulary is straightforward.
An uptrend is a sequence of higher swing highs and higher swing lows. A break of structure is price taking out the most recent swing high in an uptrend — the trend continuing. A change of character is price taking out the most recent swing low instead — the first evidence that the sequence has ended.
This is Dow theory. It was written down more than a century ago and it is one of the more durable observations in technical analysis, because it describes something structurally true: markets that are trending make progressively higher extremes, and the first failure to do so is genuinely the earliest available signal that something changed.
The new labels are not a problem. BOS and CHoCH are clearer and more precise than “the trend is intact”, and precision is worth something. What is a problem is teaching them as though they were discovered recently and are therefore secret. They are neither. Their value is that they force you to define the trend mechanically instead of eyeballing it.
On gold specifically, one caution: which timeframe you define structure on decides the answer. Gold can be making higher highs on the four-hour and lower lows on the fifteen-minute at the same moment, and both readings are correct. Structure is only meaningful relative to a stated timeframe, and a trader who switches timeframes after entry to find the reading that justifies holding a loser is using the framework to avoid a decision rather than make one.
Order blocks
The usual definition: the last opposing candle before an impulsive move away — the last down candle before a strong rally, for instance — marked as a zone that price is expected to react to when it returns.
Does this describe something real? Partly, and the honest version is more interesting than either the hype or the dismissal.
What is real: a sharp impulsive move away from a level usually means significant unfilled interest was sitting there. Whoever was buying could not fill everything before price ran. When price returns, some of that residual interest may still be waiting. The zone is a reasonable guess at where.
What is not new: this is a supply and demand zone, which is itself a refinement of support and resistance. The order block label adds a story about institutional intent that the chart cannot confirm. You cannot see who traded from a candle. You are inferring that large participants were active because the move was large, which is circular.
What is the trap: order blocks are identified after the impulsive move, which means they are always visible in hindsight and always look predictive on a historical chart. Every chart is covered in zones that worked, because the ones that did not are not labelled. This is the same hindsight problem that ruins backtests, discussed in the backtesting article, and it is far worse when the identification is manual.
The usable version: treat an order block as a supply or demand zone, with no assumption about who is in it, and require it to be tested before you trade it rather than assuming it holds.
Fair value gaps and imbalance
A fair value gap is a three-candle pattern where the middle candle moves so fast that the first candle's wick and the third candle's wick do not overlap. There is a price range that was passed through without two-sided trade.
This is the most mechanically defensible idea in the whole vocabulary, because it is purely descriptive. There is no claim about intent — it is arithmetic on three candles, and either the ranges overlap or they do not. Two traders applying it to the same gold chart will mark the same gaps.
The behavioural claim attached to it — that price tends to return and fill the gap — is testable and worth testing rather than believing. On gold it is particularly worth testing, because gold produces a great many gaps around data releases and session opens, and a fast move on genuine repricing has no obligation to come back. Some gaps fill within hours; others sit unfilled for months while the market moves on.
A useful question to answer for yourself on your own data: of gold fair value gaps formed in the Asian session, what proportion fill within the same day? Compare against the same statistic for gaps formed in the first ten minutes after a US data release. If the two numbers differ — and I would expect them to — then “gaps get filled” is not one rule, it is several, and knowing which one you are in matters more than the pattern.
Liquidity sweeps
The idea: stop-loss orders cluster in obvious places — just above a recent high, just below a recent low, above a round number. Price reaches into those clusters, triggers the stops, and then reverses.
The underlying mechanism here is straightforward and real. Stops do cluster at obvious levels, because most traders place them at the same obvious levels. A stop is a market order. A cluster of them is a pocket of guaranteed volume. Price is drawn to volume. None of this requires a conspiracy, just an understanding that resting orders are where the liquidity is.
The Asian range on gold is a textbook example: the overnight high and low are visible to everyone, stops sit beyond both, and the London open frequently reaches through one before establishing direction.
Where it goes wrong is as an explanation after the fact. Every reversal can be described as a sweep, and every continuation as “the liquidity was taken and the trend continued”. A framework that explains both outcomes equally well has explained nothing. To make it usable you have to commit in advance: which specific level, what depth beyond it counts as a sweep rather than a break, and within what time price has to reject for the idea to be valid. Those three numbers turn a narrative into a rule you can be wrong about.
How to test any of it on gold
Rather than adopting or rejecting the framework wholesale, take one component and make it mechanical:
- Define it in code or in numbers. Swing = higher than N bars either side. FVG = no overlap between bar one's wick and bar three's wick. No judgement permitted.
- State the claim as something that can fail. Not “price respects order blocks” but “after price returns to a demand zone formed by an impulse of at least X, it moves Y in my favour before moving Z against, more than half the time.”
- Count, over a fixed sample. Sixty to a hundred instances on your own broker's gold data. Count every instance, including the ones you would have talked yourself out of.
- Compare against the cost. An edge that is real but smaller than your spread plus slippage is not tradeable, which is the arithmetic in the position sizing article.
- Split by session. Almost every structural behaviour on gold differs between Asian hours and the London–New York overlap. An aggregate number hides that.
The honest summary
Market structure gives you a vocabulary for describing what price has done, and vocabulary genuinely helps — it is easier to be disciplined about a rule you can state than about a feeling. Breaks of structure, imbalance, and stop clustering all point at real mechanics.
What none of it gives you is foresight, or an explanation of who did what and why. The moment the framework is being used to narrate rather than to decide — when every outcome has a name and no outcome could have proved the reading wrong — it has stopped being analysis and become a way of feeling certain. That feeling is expensive on an instrument that moves like gold.
I trade gold, forex and index CFDs, sell short-dated option volatility, and build the MT5 and TradingView tools published on this site. Everything here is written from my own screen time and my own losses — not from a content brief. More about how I trade and write.