WhitmanTrading

What Is a Market Trend?

Market trend is a sequence of price extremes moving consistently in one direction — higher highs with higher lows, or lower highs with lower lows. It is a structural definition rather than an impression, which matters because on measured data most stretches that feel like trends are only two bars long.

Everyone knows what a trend looks like. Almost nobody can say when one started without looking backwards, and the measured data explains why.

How to read it

A price series with a sequence of higher highs and higher lows marked.
A trend is a sequence of highs and lows, not a feeling. Illustrative chart - not real market data.

An uptrend is higher highs and higher lows. Each swing peak exceeds the last, each pullback bottoms above the previous one. A downtrend is the mirror.

A steady series with each swing point labelled.
Higher highs and higher lows is the definition. Illustrative chart - not real market data.

This is a structural test, not an impression. You can check it on a chart with no indicators, and two people applying it to the same chart will usually agree — which is more than can be said for most things in technical analysis.

A rising series interrupted by a single lower low.
One lower low does not end it. Illustrative chart - not real market data.

One violation does not end a trend. A single lower low inside an uptrend is a deeper pullback, and calling it a reversal is the most common way to exit something that was working.

A falling series interrupted by a single higher high.
And one higher high does not start one. Illustrative chart - not real market data.

The reverse is equally true. One higher high in a downtrend is a bounce until the structure actually changes, and treating it as a bottom is how people buy into falling markets repeatedly.

The identification problem

A trend is always named after it exists. The second higher high is what confirms the first was part of something — so by the time the label is justified, part of the move has happened.

That is not a flaw in the method; it is the definition working. Any rule that identified trends earlier would be identifying things that have not yet met the criteria, which is a different activity called guessing.

A worked example

Here is what this site’s shared series actually did.

A choppy series with short runs in each direction.
Direction runs average 2.01 bars on this series. Illustrative chart - not real market data.

Direction runs average 2.01 bars. A “run” is consecutive bars moving the same way. The average is two. Not twenty, not eight — two.

A slow series with the longest one-way run marked.
The longest run one way was eleven bars. Illustrative chart - not real market data.

The longest run the series ever produced was 11 bars. That is the extreme, not the norm, and it happened once.

Now hold those two numbers against the experience of watching a chart. Three bars in the same direction feels like a trend establishing. On this distribution, three consecutive bars is already above average and nothing has been established at all.

A calm series where a short run resembles a trend.
So most apparent trends are two bars long. Illustrative chart - not real market data.

This is why trend-following systems lose most of their trades. They are designed to sit through many short runs that go nowhere in order to catch the rare eleven-bar one, and the arithmetic only works because the long run pays for all the short ones.

Why the timeframe decides the answer

The same price history contains several trends at once, and they can disagree. An hourly chart can be making higher highs while the daily is making lower ones, and both readings are correct because they are measuring different swing points.

That is not ambiguity to be resolved — it is the structure of the thing. A market can be in an uptrend on one horizon and a downtrend on another simultaneously, and any attempt to declare one of them the real answer is importing a preference the chart does not contain.

The practical rule is to fix the timeframe before looking. Decide which horizon you are trading, read the structure there, and accept that the others will sometimes contradict it. The failure is choosing the timeframe after seeing which one supports the position you already want.

And the shorter the timeframe, the more of what you see is noise. Direction runs average 2.01 bars regardless of the bar size, so a one-minute chart produces the same two-bar runs as a daily one — with the difference that on the one-minute chart the round-trip cost is a far larger share of what the run is worth.

The original data

On this site’s shared series, direction runs average 2.01 bars with a longest of 11. Median bar range is 0.493 and the ninetieth percentile is 1.101. A round trip costs 0.0098, about 2% of the median bar.

Put those together and the cost of chasing short runs is visible. Entering on the second bar of a run that averages two bars means entering as it ends, paying 0.0098 each time. Twenty such attempts costs about 0.196 — roughly 40% of a median bar’s entire range, spent on entries into moves that were already finished.

A candlestick chart annotated with the cost of a round trip.
And a round trip costs a share of a bar. Illustrative chart - not real market data.

The 85% figure is the other half of the picture. Of 39 twenty-bar breakouts on this series, 85% continued in the breakout direction. So trends do exist and they do persist — they are just far rarer than the impression of one.

A price series with volume shown beneath.
Volume and price are different measurements. Illustrative chart - not real market data.

When it fails

The characteristic failure is declaring a trend on two bars and sizing for eleven. The position is entered on something that matches the average run length exactly — meaning it is statistically finished — and sized as though a sustained move were underway. The run ends on schedule, the stop is hit, and the outcome is filed as bad luck. It was the measured base case, and the mistake was in the word “trend” being applied to something the data says is ordinary noise.

A falling series with a stop level marked.
A stop fills where the market is, not where you asked. Illustrative chart - not real market data.

A second failure is changing timeframe until a trend appears. There is always a timeframe on which the current move looks like one, and choosing it after the fact is fitting the frame to the answer.

A candlestick series with a gap through a level.
A gap skips the level entirely. Illustrative chart - not real market data.

A third is calling the end early. One lower low ends nothing, and acting on it exits the moves that were going to pay for everything else.

A fourth is confusing a trend with a reason. The structure describes what price has done; it contains no information about why or whether it continues.

A declining series cut short at a decision point.
Two green bars. Is that a trend? Illustrative chart - not real market data.

And a fifth is sizing the same in a two-bar run and an eleven-bar one, which treats a common event and a rare one as though they deserved equal capital.

Volatility covers how far the bars move while the trend does or does not exist. Breakout covers what the measured continuation rate actually is. And support and resistance covers the levels the swing points are measured against.

What I actually do

The number that changed how I look at charts is that direction runs on this series average two bars. Two. Everything I used to call a trend while it was happening was, on the measured distribution, a coin landing the same way twice — and I was sizing positions as though it were a regime.

— Michael Whitman

This page is educational, not financial advice. Test every idea on your own charts before risking money.