Market structure basics

Charles Dow, and why he still comes up

Charles Dow ran the Wall Street Journal at the end of the nineteenth century. He never published a method: he wrote editorials, and it was his successors, William Hamilton then Robert Rhea, who drew from them the body of principles known since as Dow theory.

A good part of that corpus has aged. Dow reasoned on two indices, industrials and railways, in an economy that no longer exists. What survived intact is the part describing how a trend is recognised, and it survived for a simple reason: it assumes nothing. No formula, no setting, no hypothesis about how participants behave. Only an observation about the succession of highs and lows.

Three of his principles still earn their keep daily. Markets move in trends of three degrees, nested like tides, waves and ripples. A trend is in force until proven otherwise, and that proof is a precise event, not an impression. Finally volume must confirm the move, failing which the move is suspect.

Those three ideas fit in a paragraph and cover most of what a beginner needs to avoid misreading the regime. The rest of this lesson unfolds them one by one.

One historical note is worth keeping, because it explains why the theory is so sober. Dow was a journalist, not a trader. He was trying to describe what the market was doing so his readers could understand it, not to sell them a system. Nothing in his writing promises a return, predicts a level or names a date. A century and a quarter later, that restraint is exactly what makes the surviving part still usable, while the methods that promised more have all been dated by the decade that produced them.

An uptrend: higher lows

This is the most useful definition in the whole programme, and it fits in one sentence: an uptrend is a succession of higher highs and higher lows. Two series rising together, nothing more.

The order of the two matters. It is the lows that carry the information, not the highs. A market can print a new high in a burst of panic and fall straight back: the high says nothing about willingness to buy. A low higher than the previous one says something far more solid, namely that buyers stepped in earlier than last time, without waiting for last time's price.

It is also why a pullback is an opportunity in a trend and a trap in a range. In an uptrend, the pullback carries price towards somewhere buyers have already shown they act, and they tend to act higher each time.

One exercise worth more than ten videos: open a daily chart, mark the last four lows by hand, and write down whether they rise, fall or line up. Three minutes, and you know your regime. Most bad decisions come from nobody having spent those three minutes.

A downtrend: lower highs

The exact mirror. A downtrend is a succession of lower highs and lower lows. This time the highs carry the information: each bounce runs out of steam earlier than the last, meaning sellers step in at a progressively lower price, without waiting for last time's level.

One asymmetry is worth knowing, because it surprises those discovering short selling. Falls are on average faster and more vertical than rises. A position taken with the fall therefore moves faster, in both directions, and a sizing error is paid for more quickly.

The other classic trap of a downtrend is the violent bounce. A falling market regularly produces spectacular recoveries, which read as reversals and are not, as long as the last high stands. Structure settles that argument in ten seconds, where the impression can last for days.

The range, or the absence of structure

When neither series rises nor falls, there is no trend. The highs line up roughly, so do the lows, and price travels between them. That is a range, and it is the state markets spend most of their time in.

It is not an anomaly nor a pause: it is the normal regime. The auction theory lesson explains why, and it says the same thing in another language. A market searches for a price where business gets done, and once it finds it, it stays.

What to take from it is practical. In a range, breaking a boundary is usually a trap, and the opportunity sits at the boundaries, not in the middle. Applying trend logic to a range produces a run of losing trades none of which is an analysis error: they are all the same regime error.

If you cannot say within ten seconds which of the three regimes you are looking at, there probably is none, and doing nothing is a valid answer. It is often the most profitable answer available that day.

The turn: two conditions, in order

A trend does not stop because it is “tired”, because it “went up a lot”, or because an indicator shows an extreme reading. Those three reasons are expensive and are not reasons.

It stops when two things happen, in this order. First price stops making new extremes in its direction: a high fails to exceed the previous one. Then it breaks the last low that supported the series. As long as that low holds, a trend that looks winded is a trend that continues.

It is the most useful reference in the lesson, because it is checkable: you can point at it on the chart, and two people applying it to the same chart agree. An impression of tiredness, no.

⚠️ One clarification that prevents many premature entries: breaking that low announces the end of the previous trend, not the start of an opposite one. What follows is very often a range. Confusing the two makes people short at the exact moment a market merely stops rising, which is a position with no edge.

This is the modern translation of Dow's principle that a trend stays in force until proven otherwise. What structure reading contributed was defining what that proof is.

Confirmation by volume

Dow held volume to be a judge, and it is the principle easiest to verify today. The reasoning is direct: a move carried by many transactions mobilised many participants, a move without volume mobilised nobody.

In a healthy uptrend, volume grows with the trend and contracts during pullbacks. The opposite is a warning: when pullbacks come on rising volume and rallies on thinning volume, the structure still holds but participation is leaving it.

On a break the question becomes blunt. A break on volume well above normal says the level was genuinely contested. The same break without volume mostly says nobody was there, and that is the signature of the false break, the one back inside the zone two candles later.

⛔ A warning that holds across the programme: this reading needs REAL volume, therefore a centralised market. On the spot currency market there is no single book, and the volume your chart displays counts price changes your broker saw. That is a measure of activity, not volume traded, and basing a confirmation on it means confirming one thing with another.

If you trade currencies and want a substitute, the closest honest one is the volume of the corresponding currency futures contract, which is centralised and published. It does not cover the whole market, far from it, but what it shows was genuinely dealt. That is already a great deal more than a tick count from one broker among hundreds.

What structure does not tell you

It gives no target. Knowing a trend is under way says nothing about how far it goes, and nothing in Dow theory claims otherwise. Targets are set with something else: the zones from the price action lesson, or the references from auction theory.

It gives no timing. A higher low can form in an hour or in three weeks. Structure describes a state, not a calendar, which is why it pairs with a finer trigger rather than serving alone.

Finally it depends entirely on timeframe. An uptrend on the daily contains dozens of downtrends on the five-minute, all perfectly real. Two people arguing about a market's direction are almost always looking at two different degrees, and Dow had already written it, talking about tides, waves and ripples.

The cure is the one from the technical analysis lesson: pick your decision timeframe, keep it, and never drop a level to justify a position that hurts.

The three most common reading errors

The first is reading structure on the wrong timeframe. It is so common it deserves a name: the decision is made on the daily, the entry is taken on the five-minute, and the position is then judged on the five-minute. The structure that motivated the entry is never the one that decides the exit, and the trade loses all its logic.

The second is counting highs and lows without qualifying them. They are not all equal. A low formed on a thirty-second wick during a data release is not a structural low: nobody had time to take a decision there. A low where price stayed for several candles is one. The practical test is simple: if you have to zoom in to see it, it does not count.

The third is declaring a turn on the first candle that breaks. A break is judged at the close, for the same reason a candle is only read once it has shut: price dips below a low and comes back above it dozens of times a week, and that is exactly what the liquidity lesson describes.

Those three errors share one thing: they all conclude faster than the chart allows. Structure is a slow tool, and that is its main quality.

Practising: the four-low survey

Structure is not learnt by reading, it is learnt by surveying. Here is the shortest exercise that produces real progress, and it takes ten minutes a day for two weeks.

Open the daily chart of the instrument you chose. Mark the last four lows and the last four highs by hand. Write one line: they rise, they fall, or they line up. Then note your regime, up, down or range.

Then drop to your decision timeframe and run exactly the same survey. Note whether the two regimes agree. That last question is the one that pays: when both degrees are aligned, setups taken in their direction do markedly better, and this is not a belief to take on trust, it is something you will measure on your own survey.

After two weeks you will have a dozen surveys, and above all you will no longer be able to open a chart without seeing both series. That is the point where reading stops being a conscious effort, and it is the whole objective of this lesson.

Keep those surveys. Compared against your trades for the month, they answer a question very few beginners know how to ask: do I mostly lose when the two degrees contradict each other? If the answer is yes, you have just found a filter that costs nothing and removes a share of your worst trades from the record.