How to Identify a Trend: Market Structure in Highs and Lows

A trend is defined by highs and lows

The most basic way to judge a trend on a chart is the position of highs and lows. As price swings up and down, if each high is higher than the last and each low is higher than the last, it is called an uptrend; if highs and lows both keep getting lower, it is a downtrend. If they do not stack in one direction, such as highs rising while lows fall, or both staying around the same level, it is treated as sideways or directionless. Participants' behavior explains why this shape forms: in an uptrend, every pullback meets buyers who step in above the previous low, so the lows rise like stairs. Because this standard uses only price, it works without any indicator, and it is easy to explain to someone else what you looked at to reach your view. The starting point is to compare at least two highs and two lows, four turning points in all. With only three, it is too early to call it a trend; "forming" is the more accurate word.

• Uptrend: higher highs and higher lows
• Downtrend: lower highs and lower lows
• Anything else: sideways or no clear direction

How to pick swing highs and lows

To compare highs and lows, you first have to decide which bar counts as a high. A common definition is the swing. If a bar's high is above the highs of the N bars to its left and the N bars to its right, it is a swing high; if its low is below the lows of the N bars on each side, it is a swing low. A small N catches even minor wiggles as highs and lows, so the structure changes often; a large N leaves only the peaks and troughs of big waves. This N is what the support and resistance auto-detection tool calls "swing strength," and you can choose 3, 5 or 8 (default 5). Changing N on the same chart can change the trend call. How a neighboring bar with an equal value is handled also varies slightly between tools. A common mistake is changing N whenever you dislike the result; you end up choosing the yardstick that fits the conclusion you wanted. It is better to fix one N per timeframe and write down the reason whenever you change it.

Swings are always confirmed late

The swing definition includes the N bars to the right. So you can only know that a bar is a swing high after N more bars have closed and no higher high appeared in between. With N set to 5 on the daily chart, a high is confirmed only five days after it printed, and by then price may already be well below it. This is why highs and lows look crisp on a past chart, while in real time there is only "a bar that might be a high." This delay is not a flaw in the definition but its price. Shrink N to confirm sooner and highs change with every small wiggle; whatever delay you remove comes back as false highs. Looking at swings on a past chart and thinking "selling there would have worked" leaves out that the information did not exist in real time. Make a habit of checking when each swing the tool marks was actually confirmed.

Where the structure breaks

When reading an uptrend's structure, the reference price is the most recent swing low. A series of higher lows is half of the uptrend structure, so a close below that low breaks the "higher low" condition. This price draws attention because many people may have bought at that low and placed their stop-losses just under it; if price goes below, those orders can fire together and enlarge the move. A wick that briefly dips below and comes back is treated separately, and that difference is covered in the false breakout article. A broken structure does not instantly make a downtrend. To call it a downtrend, highs and lows need to keep falling together, and until then the definition fits better if you treat the direction as unclear. In a downtrend, flip everything: a close above the most recent swing high plays the same role.

Each timeframe has its own trend

At the same moment it is common for the daily chart to show an uptrend while the hourly chart shows a downtrend. A move that is just one pullback on the daily can look, on the hourly, like a small downtrend with highs and lows falling for days. Neither is wrong; they are looking at flows of different sizes, so when you talk about a trend, state which timeframe you mean to make it clear. Different participants also watch different timeframes: to someone who entered with a horizon of weeks, a one-day drop is noise, while to someone watching hours it is the trend. A common order of checks is to set the big picture on the longer timeframes first, then step down to shorter ones to see where things stand within it. Reading several timeframes together is covered in the multi-timeframe article, and the multi-timeframe trend table tool gathers the trend from 15-minute up to weekly bars in one table.

Indicators summarize the same structure in different ways

Most indicators said to show trend are simply other ways of summarizing the flow of highs and lows. A rising moving average means the average of recent prices is higher than before, and Supertrend sets direction by whether the close crosses a trailing line placed some multiple of ATR (average true range) away. Indicators have fixed rules, so anyone who calculates them gets the same answer, but because they are computed from past prices, they report a change of direction late, just like swings. A common illusion is to treat several indicators pointing the same way as that much more evidence. They are summaries of the same prices and usually move together, so three agreeing are not three independent reasons. In sideways ranges, these indicators also tend to flip back and forth and produce a string of late signals. When an indicator's direction disagrees with the high-low structure, before arguing which is right, check whether the two are looking at different periods.

• Moving average slope: if the average is rising, the bias is up
• Supertrend: direction flips when the close crosses the trailing line
• Multi-timeframe trend table: a score from −3 to +3 based on where the close sits relative to EMA20 and EMA50, plus the EMA20 slope

Limits: judging a trend is not forecasting

A judgment based on highs and lows describes the shape so far; it does not tell you which way price will go next. When this course measured the daily bars of 10 coins on Binance (each from its Binance listing date to September 2026), of all 29,746 bars that had a close 20 bars later to compare with, the close 20 bars later was higher 50.4% of the time. The average return 20 bars later was +4.62%, but that figure was pulled up by a small number of large rallies; the median was +0.15%. This was not a test of trending stretches only, but it does mean that even over a period when prices rose a great deal, picking any day at random and looking 20 days ahead showed a gain only a little more than half the time. A gap this wide between mean and median also means that, apart from a few big moves, most days went almost nowhere. A trend judgment is a framework for deciding where to put a stop or which rules to use, not a guarantee that the trend will continue. Using it to ask which kind of rule suits the current environment, rather than expecting it to raise your hit rate, fits the measurement better.

What looks different in crypto and stocks

Crypto trades 24 hours a day with no gaps between bars, so highs and lows form on a continuous price path. With stocks, news that arrives while the market is closed is priced in all at once as a gap at the next open, so instead of drifting below a swing low during the session, price can open straight past it. In that case there was no chance to trade at the price where the structure broke, which is a real difference from crypto. Korean stocks have a daily price limit of 30% up or down, capping one day's move, while US stocks have no daily limit on individual names but do have mechanisms that briefly halt trading during sharp moves. Volatility differs too. Daily swings of several percent are common in crypto, so swings form often even with the same N, whereas US indices and large tech stocks, and large Korean names such as Samsung Electronics and SK Hynix, tend to swing less, so the same N catches swings at different intervals and sizes. Before carrying an N from crypto over to stocks, look at how swings actually print on the chart. Also remember that stock daily charts have no bars on weekends and holidays, so "20 bars" spans a longer calendar period than in crypto.

• Crypto: continuous 24 hours, with almost no gaps
• Stocks: news between sessions shows up as an opening gap
• Korean stocks have a daily price limit
• Twenty daily bars in stocks cover a longer calendar period than in crypto

Reading it on a live chart

The most confusing thing on a live chart is the rightmost bar still in progress. Its high and low keep changing until it closes, so an in-progress bar dipping below the latest swing low does not mean the structure has broken. That is exactly why structure is judged on the close: it is not rare for price to recover before the close and leave only a wick. Swing markers work the same way. A swing high the tool has just marked was confirmed as the N bars to its right closed, and a bar that looked like a "high candidate" before that disappears once a new high prints. Indicators computed from the close, such as Supertrend or the trend table score, also change with every tick on the in-progress bar, so a change of direction can appear and vanish repeatedly. So read signals on closed bars, and treat in-progress values only as "if the bar ended right now."

Common mistakes

The mistakes that keep recurring in trend reading usually come from blurring the standard. The items below help you keep the procedure, more than the judgment itself. Writing down the timeframe, swing strength and whether the bar has closed before calling a trend filters out most of them. When reading someone else's analysis, check first whether those three are stated.

• Calling something an "uptrend" without saying which timeframe
• Treating a single wick below a low as a broken structure
• Using an unconfirmed swing as if it were confirmed
• Counting several agreeing indicators as independent evidence
• Taking a trend judgment as a forecast of the next direction