Market Structure in Trading: Highs, Lows, BOS and CHoCH
Michael Whitman
·
Trading since 2017, full-time since 2020
·
YouTube creator, 137K subscribers
·
Published 2026-09-05
·
Updated
·
Reviewed
Market structure is the pattern of highs and lows on a chart. An uptrend makes higher highs and higher lows; a downtrend makes lower highs and lower lows. It matters because it gives you one specific price at which your read of the market is wrong.
Covered on this page: TradingView and LuxAlgo.
Market structure is the pattern of swing highs and swing lows on a chart: higher highs and higher lows make an uptrend, lower highs and lower lows a downtrend, and anything else is a range. On SPY since 1993, five-bar daily changes of character down came a median 4.28% below the top of the leg.
It is not a drawing or an indicator but a description of what price has already done, and it names the one price, the last swing low or high, where the current read becomes wrong.
An uptrend, defined exactly
An uptrend is a sequence where each high is above the previous high, and each low is above the previous low. Both conditions. Not one. The first half has its own page on the higher high.
That second condition does most of the work. A collapse can print a higher high; it cannot keep printing higher lows.
A downtrend is the mirror
Lower highs and lower lows: the same rule inverted, with the full treatment on the downtrend page.
And a range is neither
If the highs are roughly level and the lows are roughly level, there is no trend. There is a range.
This is the state most often mislabeled, as the failure section below explains.
Break of structure vs change of character
These two get used interchangeably. They are opposites, and confusing them is expensive.
A break of structure is price taking out the previous high in an uptrend. The trend continued. It is confirmation, not a signal to do anything new.
A change of character is price taking out the last higher low in that same uptrend. The trend broke. The pattern that defined it no longer holds.
The simplest way to keep them straight: a break of structure confirms what you thought. A change of character tells you that you were wrong. One is agreement, the other is news.
Internal structure vs swing structure
Swing structure is the big picture, the highs and lows you would still see zoomed out; internal structure is the smaller moves inside a swing. Most tools show both at once. An internal break inside a swing means something different from one on its own, but build the read around the swing, or moves that were never structure will stop you out.
A wick through is not a break
Price traded above the level. It did not close above it, so under a close-based rule nothing has broken yet. Nor is the wick evidence of a reversal: on QQQ since 1999, 76 of 96 wick-only breaches of a swing high closed above it within ten sessions, as the liquidity sweep vs break of structure comparison shows. Waiting for the close costs a few points of entry and keeps the rule mechanical.
Which swings actually count
The commonest mistake is marking every small wiggle.
A swing point worth marking is one where price turned and went somewhere. If a high was exceeded two candles later, it was noise. The swing high and swing low page covers how the points themselves are defined.
The practical test: if you zoom out one timeframe and the point disappears, it was never structure.
How often structure really changes
Exactly as often as your swing rule lets it. On the same SPY prices, the smallest common swing rule flips the read about every three weeks and a wide weekly rule about every nine months, and the data section shows the whole range. So the swing width is a decision made before the chart is open; changing it after a loss fits the rule to the trade you just lost.
On a daily chart, expect a flip about as often as a continuation. The textbook picture is a long run of breaks of structure and one change of character at the end. On SPY’s daily bars the two came in roughly equal numbers, which is why a single daily break is weak evidence that the trend will carry on. A market structure shift is the same event under another name.
A worked example
Take SPY’s daily chart from July to September 2026 and use one fixed rule: a swing high is a day whose high is above the highs of the five sessions before it and the five after it, and a swing low is the mirror. Because it needs five later sessions, a swing is only known a week after it forms.
The read starts down. A close below a swing low on 10 June 2026 turned it down, and another on 23 July confirmed it.
3 August: change of character up. SPY’s high of $755.58 on 15 July had become a confirmed swing high on 22 July. On 3 August SPY closed at $757.67, above it. The lower-high pattern was broken, so the read flipped to up.
13 August: break of structure up. The 5 August high of $776.85, known from 12 August, was taken out by a close of $777.88. That confirms the up read and asks nothing new of you.
The line that mattered. The 20 August low of $762.04 became a confirmed higher low on 27 August. From then on there was a specific number: above it, the up read held. A close below it and the sequence was broken — not “looking weak,” broken, by the same definition that made it a trend.
1 September: change of character down. SPY closed at $761.78, 26 cents under $762.04. Then on 10 September it closed at $757.83, below the 1 September swing low of $759.48, a break of structure that confirmed the new down read.
That number existed three sessions before you needed it, which is the whole point of marking structure at all. It also came late: the 1 September close was 2.26% below the 13 August high of $779.37, 13 sessions after it. And the 26-cent margin is a reminder that the rule is mechanical: a wider swing rule would have drawn a different line and given a different answer on the same day.
The original data
The sample. SPY daily bars from 29 January 1993 to 25 September 2026, 8,472 sessions or 33.7 years, from Yahoo Finance.
The rule. It is the one in the worked example with the width changed: a swing needs a high (or low) beyond that many bars on each side, each swing is known only after those later bars close, and a close beyond the latest known swing breaks it. A newer swing replaces the older one on the day it becomes known, before that day’s close is judged, even if the older one was never broken. A break in the direction of the current read counts as a break of structure; a break against it counts as a change of character. Weekly bars are built from the same days. The daily five-bar events are listed in the SPY structure events file.
| Chart and swing width | Swing points a year | Breaks of structure a year | Changes of character a year | Median gap between changes |
|---|---|---|---|---|
| Daily, 2 bars each side | 65.6 | 13.4 | 12.9 | 21 days |
| Daily, 5 bars | 29.1 | 5.9 | 6.5 | 41 days |
| Daily, 10 bars | 13.7 | 2.8 | 3.0 | 78 days |
| Weekly, 2 bars | 12.1 | 2.6 | 1.9 | 105 days |
| Weekly, 5 bars | 5.7 | 1.5 | 1.1 | 266 days |
The swing width decided the count, not the market. The same 33.7 years of prices gave 12.9 changes of character a year with two-bar daily swings, 3.0 with ten-bar swings and 1.1 with five-bar weekly swings. The gaps are calendar days.
Daily breaks split evenly between continuing and flipping. With five-bar daily swings there were 199 breaks of structure and 220 changes of character, so only 47.5% of breaks confirmed the existing read. On weekly bars the share of confirmations rose to 57.6% with two-bar swings and 59.1% with five-bar swings.
The two-bar rule is the one behind the SPY study on the internal and external liquidity page.
How late a change of character arrives. A fresh download of the same SPY bars, now running to 23 October 2026 (8,492 sessions, 1,761 weeks), was put through the same rules. For every change of character, the leg it ended runs from the day the read turned to the day it flipped; for a flip down, the lag is how far the confirming close sat below the leg’s highest high, and how many bars after that high it came. A flip up is the mirror, measured from the leg’s lowest low.
| Chart and swing width | Flips down | Close below the leg’s high | Bars after the high | Flips up | Close above the leg’s low | Bars after the low |
|---|---|---|---|---|---|---|
| Daily, 2 bars | 219 | 3.06% | 5 | 219 | 4.35% | 6 |
| Daily, 5 bars | 110 | 4.28% | 9 | 111 | 6.12% | 14 |
| Daily, 10 bars | 50 | 5.88% | 18 | 51 | 8.64% | 23 |
| Weekly, 2 bars | 32 | 6.74% | 5 weeks | 32 | 11.61% | 5 weeks |
| Weekly, 5 bars | 18 | 8.70% | 9.5 weeks | 18 | 17.96% | 14 weeks |
The finding: the wider the swing rule, the more of the move was gone before it confirmed. With two-bar daily swings, the close that flipped the read down sat a median 3.06% below the top, 5 sessions after it. With five-bar swings it was 4.28% and 9 sessions, and in 43 of the 110 flips down the confirming close was 5% or more below the top. Five-bar weekly swings gave fewer, slower signals: 18 flips down, a median 8.70% under the top. Flips up sat further from the leg’s extreme on every width. The largest daily five-bar lag was 14 April 2000, a close of $136.00 against the 24 March high of $155.75, 12.68% lower. Every flip is in the change of character file and the medians in the lag summary.
Checked against Nasdaq. Run on Nasdaq’s SPY prices for the 2,514 sessions from 24 October 2016 to 23 October 2026, the five-bar daily rule found the same 128 breaks of structure and changes of character on the same dates as Yahoo’s bars, including the same 63 changes of character, with the same median lag of 4.85%.
The video record. Across our study of 24,971 trading videos, 319 have market structure in the title. The median one gets 11,735 views, 65% never pass 50,000, and the median length is 19.6 minutes. The corpus carries description text for 93 of those 319, and across those 93, 4 mention invalidation, failure, or getting it wrong.
Small window — 93 videos, not 319 — so treat it as a hint rather than a census.
How this connects
It is the frame nearly every other smart money concept hangs on. An order block, a fair value gap and a liquidity sweep are all defined relative to it. Support and resistance is the older vocabulary for a similar idea: structure is about sequence, levels are about price.
When it fails
You marked structure on the wrong timeframe
A five-minute chart and a daily chart routinely disagree, and both are correct; the lower-timeframe break that stopped you out was noise on the chart you were trading.
Pick the timeframe you intend to trade, mark structure there, and treat lower timeframes as detail rather than as contradiction. Ranking the timeframes first is the idea behind top-down analysis.
You needed the trend to be there
A range with a slight upward tilt becomes an uptrend the moment you have a reason to want one. The honest check is whether you would mark the same structure on a chart with the ticker hidden.
You marked it after the move
Scrolling back to find the structure that “worked” is trivial and feels like analysis.
What structure is actually for
Not prediction. Structure gives you the one price at which your read is wrong. If you are long in an uptrend, the last higher low is that price: visible, specific, and decided before you needed it.
Questions people ask
What is market structure in trading?
The sequence of swing highs and swing lows on a chart. Higher highs and higher lows make an uptrend, lower highs and lower lows a downtrend, and anything else is a range.
What is the difference between a break of structure and a change of character?
A break of structure takes out a swing level in the direction of the current read and confirms it. A change of character takes out the level against it, such as the last higher low in an uptrend, and flips the read.
How late does a change of character confirm?
Later as the swing rule widens. On SPY since 1993, a five-bar daily flip down closed a median 4.28% below the top of the leg, 9 sessions after it; with five-bar weekly swings it was 8.70%.
How many bars make a swing high?
No single number; it is a rule you choose. On SPY since 1993, two-bar daily swings flipped the read 12.9 times a year and ten-bar swings 3.0 times.
Related
Once you can read structure, an order block becomes readable, because the break of structure is the thing that validates one.
The gap a fast move leaves behind is a fair value gap, and it usually sits right where structure broke.
And the reason price returns to these levels at all is liquidity — orders resting where everyone can see they should be.
To practice the reading on a live chart, how to read market structure walks through it in steps.
The setting most people get wrong is showing every internal structure break at once. Zoom out on a daily chart and there really are not many genuine breaks - a handful across months. If your chart is covered in them, the indicator is showing you noise, not structure, and you will end up trading the noise.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money. Some links on this page earn a commission if you buy through them. It costs you nothing and it does not decide what appears here or in what order — how these pages are made is set out in our methodology.