Ask ten traders what an uptrend is and most will say “price going up”, which is not a definition, it is a description of the last thing they saw. The usable definition is older and stricter: higher highs and higher lows. It is short enough to memorise and precise enough to argue with, and the arguing is the useful part, because it forces you to say exactly which highs and which lows you mean.
An uptrend is a sequence of higher swing highs and higher swing lows. A downtrend is the mirror: lower swing highs and lower swing lows. Anything that fails both tests is a range. A swing high is a candle whose high exceeds the highs on both sides of it over a chosen lookback, so the definition is timeframe-dependent by construction.
Everything below unpacks that paragraph: what a swing point actually is, why the lookback is a choice you are making whether you admit it or not, what the end of a trend looks like in the same vocabulary, and why this structure is the “prior trend” that every reversal candlestick pattern quietly depends on.
What counts as a swing high, and why the lookback is a choice
Bulkowski’s glossary defines a minor high as a significant turning point that results in a peak, and a minor low as the same thing resulting in a valley. That is the idea, but “significant” is doing a lot of work. To make it checkable you need a rule, and the common one is this: a candle is a swing high if its high is above the highs of the N candles on each side of it. N is the lookback. With N of 1 you get a swing point at almost every wiggle. With N of 10 you get only the turns that held for two trading weeks on either side.
There is no correct N. There is only the N you have chosen, and the trend you see is a function of it. This is not a flaw in the definition, it is the definition being honest about something every chart reader does anyway: a trend is a statement about a sequence of turning points, and turning points only exist relative to a scale. Fix the scale first, then read the structure. Read the structure first and you will pick the scale that agrees with what you wanted to see.
A worked sequence, with illustrative prices
Take a stock on a daily chart, illustrative prices throughout. It bottoms at 40 and rallies to 50. That 50 is a swing high once price pulls back from it. The pullback holds at 45 and turns up, so 45 is a swing low, and it is higher than 40. Price then runs to 56: a higher high. The next pullback stops at 51: a higher low. So far the sequence is 40, 50, 45, 56, 51, and the trend definition is satisfied twice over.
Now price rallies from 51 and stalls at 54. That is the first lower high. On its own, a lower high establishes very little. Every uptrend contains pullbacks, and a failed attempt at a new high can simply mean the next higher low is forming a little further down. Structurally, though, one of the two conditions has now failed, and the trend is on notice.
Then price falls through 51 and closes at 49. That is the first lower low, and it is the event that changes the classification. The sequence now reads 56, 51, 54, 49: a lower high followed by a lower low, which is the downtrend definition satisfied once. It is not a downtrend yet in any confident sense; a single pair is the minimum evidence, not a verdict. But it is no longer an uptrend by the only definition that matters, and that is worth being clear about before you decide what the chart is doing.
Bulkowski’s 1-2-3 trend change method formalises almost exactly this reading in three steps: price closes through the trendline, price tests the prior extreme, and price then closes beyond the intervening low or high. His caveat is worth carrying too: how much of the chart you display determines which trendline you draw, and that can change the answer dramatically. Same lookback problem, different clothes.
Which comes first, the lower high or the lower low
Traders argue about this and the answer is that they do different jobs. The lower high is the earliest observable failure, and it is also the most common false alarm, because ordinary pullbacks produce them constantly. The lower low is the later, stronger observation: it means the price where buyers last stepped in did not hold, which is the thing an uptrend cannot survive.
So a useful habit is to treat the first lower high as a question and the first lower low as the answer. Between the two, the chart is ambiguous and you should say so, rather than resolving the ambiguity with hope.
Why the daily and the hourly can disagree without contradiction
The same instrument is frequently in an uptrend on the daily chart and a downtrend on the hourly. Beginners read this as one of the charts being wrong. Neither is. Dow theory described the market as having three movements at once, a primary movement lasting from under a year to several years, a secondary reaction lasting roughly ten days to three months, and a short swing of hours to a month or more, and noted that the three can run simultaneously. In structural terms, the hourly chart’s lower highs and lower lows are simply the internal anatomy of one daily higher low that has not finished forming yet.
The practical rule falls out of the definition. Decide which timeframe you are reading before you look, because that fixes the lookback, which fixes the swing points, which fixes the trend. If you need the mechanics of how candles aggregate across timeframes, how to read a stock chart covers it.
Trend structure is the prior trend every reversal pattern needs
Here is the connection that makes this more than vocabulary. Every reversal candlestick pattern carries a context requirement: it only means something after a trend in the opposite direction. “After a decline” is not a vibe. In these terms it means the chart has been printing lower highs and lower lows, and the pattern is forming at or below the last swing low.
The structure also supplies the invalidation, which is the part most pattern lists leave blank. A bullish reversal read after a decline is wrong on a close below the last swing low, because that close is, by definition, a new lower low, and the downtrend has simply continued. The candle did not fail mysteriously; the structure told you exactly what failure would look like before it happened. The reversal candlestick patterns post takes each pattern through that logic.
Chart patterns are the same idea at a larger scale. A triple top is three attempts at a higher high that all failed at roughly the same level, and the pattern is only confirmed on a close below the swing low between them, which is the first lower low. The triple top post walks through that. And the levels where swing lows have repeatedly held are what most people mean when they say support, which is why support and resistance is best drawn from swing points rather than from wherever a line happens to fit. The same rule applies to a rising line under higher lows, which is the subject of how to draw trend lines.
What this framework does not do
It does not predict. A sequence of higher highs and higher lows is a classification of what has already printed, and the moment it stops being true is only knowable after the close that broke it. Bulkowski’s own note on up trendlines makes the same point from the other side: you know the trend has changed when the trendline is pierced, which is to say after the fact. Any tool that claims to know the trend has ended before a lower low has printed is guessing, and calling the guess structure.
What the framework does do is turn “I think it is going up” into three checkable statements: which timeframe, which swing points, and which price would break the sequence. That is the whole discipline, and it is enough.
For the patterns that sit inside this structure, each drawn with its context requirement and invalidation, see the candlestick pattern library. If the four prices in a candle are still new to you, how to read a stock chart is where to start.
Sources
Every figure above was checked against these before publishing.
- Wikipedia, Market trend
- Wikipedia, Dow theory
- Bulkowski, Up trendlines
- Bulkowski, 1-2-3 trend change
- Bulkowski, Glossary (minor high, minor low)