Strip the indicators off a chart and you are left with a sequence of turning points. Price runs, stalls, pulls back, runs again. Each stall leaves a high or a low on the chart, and the relationship between consecutive stalls is the whole of market structure. Higher highs with higher lows means buyers keep paying more and sellers keep defending less. Lower highs with lower lows means the opposite. Everything else in the vocabulary is built on those two patterns.
The reason to bother with it is not prediction. It is that structure gives you a specific price at which your read is wrong, which is the only thing a stop loss can be built from.
Defining a swing point so it stops moving
Most arguments about structure are actually arguments about which highs count. A swing point needs a mechanical definition, and any consistent one works as long as you apply it every time.
- A swing high is a candle whose high is above the highs of the candles either side of it, using a fixed lookback, commonly two or three bars each way.
- A swing low is the mirror: a candle with a lower low than its neighbours across the same lookback.
- The swing is only confirmed once the required number of bars has printed after it. Until then you are looking at a candidate, not a level.
- A move that fails to produce a new swing point in the trend direction is a pause. It is not yet a change in the sequence.
That confirmation lag is where most of the frustration lives. Structure is always identified after the fact, and traders who cannot accept that keep marking swing points on the live candle and redrawing them minutes later. Fix the lookback, fix the rule, accept the delay.
A break of structure versus a failed push
A break of structure is a close beyond the most recent confirmed swing point in the trend direction. In an uptrend that is a close above the prior swing high. The break says the last group of sellers who defended that level did not hold it.
The word close is doing a lot of work. A wick that pierces the swing high by four pips and closes back underneath it is a failed push, and it often means the opposite of a break: someone sold into the liquidity sitting above the level. Traders who treat wicks as breaks get chopped apart in ranges, which is the most common way this framework fails in practice. The distinction matters enough that it is worth reading alongside how support and resistance levels actually behave when they are tested repeatedly.
The more interesting event is the first break against the trend. An uptrend that has printed four higher lows and then closes below the most recent one has not necessarily reversed, but it has stopped doing the thing that defined it. Some traders call that a change of character. It is a signal to stop adding, not a signal to flip direction.
Timeframes disagree on purpose
The daily chart can be in a clean uptrend while the fifteen minute chart prints three lower highs in a row. Both readings are correct. The fifteen minute sequence is describing a pullback that the daily chart has not finished registering.
Decide in advance which chart holds the bias, and write it down. A trader who takes direction from the four hour chart and then reverses on a five minute break of structure has no method, only a running argument with themselves. Choosing a timeframe set is a structural decision, not a preference.
A workable arrangement is three charts with fixed jobs: a higher timeframe for direction, a middle one to mark the swing points and the levels, and a lower one purely for entry timing. The lower chart never overrules the higher one. It only answers the question of when, once the higher chart has answered whether.
Ranges have structure too
When the sequence stops making progress in either direction, you get equal highs and equal lows instead of a staircase. This is not the absence of structure, it is a different one, and it is where most of the trading week actually happens. Ranges have a top, a bottom and a midpoint, and price spends its time rotating between them until one boundary produces a close outside it that holds.
The transition is what pays. Extended ranges build clusters of stop orders just beyond both boundaries, which is exactly why the first move out of a range so often reverses: the run was aimed at those orders rather than at a new price level. Wyckoff's accumulation and distribution schematics are an older, more detailed language for the same phenomenon, and the spring or upthrust in that framework is the same event as a failed break in this one.
Where the reading breaks down
Structure fails hardest in three conditions. Around scheduled data releases, price can trade through two or three swing points in ninety seconds, which makes every level on the chart briefly meaningless. In thin holiday sessions, a single large order can print a swing point that no one will ever defend again. And on illiquid instruments, the swing sequence describes one participant's order flow rather than a market's.
The deeper problem is confirmation bias. Once a trader has decided the market is bullish, marginal swing points get promoted or ignored to keep the story intact. The defence is the mechanical definition from the first section, applied before you have a position. If the lookback is fixed, the chart marks itself and you get to disagree with it rather than with yourself.
Turning it into a rule
Structure earns its place when it produces a level, an invalidation and a size. In an uptrend, the last confirmed higher low is the price at which the sequence would break, which makes it the natural place for the stop. The distance between entry and that low sets the position size, and the next unbroken swing high above sets a first target. That chain, level to invalidation to size, is what makes structure usable rather than decorative, and it works the same way inside a trend following approach as it does for a single discretionary trade.
What it cannot do is tell you what happens next. It describes the last few exchanges between buyers and sellers, and it is silent about the next one. Trading on any framework carries a high risk of loss, and a break of structure is a reason to define risk, never a reason to skip defining it.
"Structure tells you which side has been winning, not which side wins next. People treat a break of structure like a signal. It is a scoreboard."
— Alex Onta, Executive Director, SINGUARD
Key Takeaways
- Fix a mechanical swing point definition with a set lookback, and accept that every confirmation arrives a few bars late.
- A break of structure is a close beyond the level; a wick that closes back inside is a failed push and often means the reverse.
- Timeframes are meant to disagree, so name the chart that holds the bias before the session starts.
- The last confirmed swing point in the trend direction is an invalidation price, which is what turns structure into a position size.
Frequently Asked Questions
What counts as a break of market structure?
A break of structure happens when price closes beyond the most recent confirmed swing point in the direction of the existing trend, for example a close above the last swing high in an uptrend. The word close matters. A wick through the level that closes back inside is a failed push, not a break, and traders who accept wicks as breaks get stopped out repeatedly in ranges.
Which timeframe should market structure be read on?
Structure exists on every timeframe and the answers disagree by design. A common approach is to set direction on the daily or four hour chart, mark the swing points there, and use a lower timeframe only for entry timing. What matters is deciding in advance which chart holds the bias so that a lower timeframe move against you is not mistaken for a reversal.
Does market structure predict where price is going?
No. Structure is a record of which side has been winning the recent exchanges. It describes what has already happened and gives a defined level to be wrong at. Every structure reading is confirmed after the fact, so it should be used to place stops and define invalidation rather than to forecast. Trading on any method carries a high risk of loss.