Elliott Wave Theory

What Elliott Wave Theory Actually Says

A plain account of the structure, the rules, and why most counts you see are wrong.

Where Elliott Wave theory comes from

An American accountant named Ralph Nelson Elliott set it out in 1938. In the early 1930s, recovering from an illness that had ended his career, he did what an accountant does with an unwelcome amount of time: he went through the numbers. Seventy-five years of stock market data, at every interval he could find, hour by hour in places.

What he reported was not a trading method. It was an observation about shape. Prices did not wander. They moved in a repeating sequence — a push, a pull back, a longer push — and the same sequence appeared whether he looked at a decade or an afternoon.

He was not the first to notice that markets have rhythm. He was the first to write down the rules that a rhythm has to obey.

Why the pattern exists at all

The structure is not arithmetic. It comes from people.

A market is a large number of participants changing their minds at different speeds. The first move up is made by the few who see something early. The pull back is made by everyone who does not believe it. The long third move is what happens when the doubters change sides, and the last move up is made by the people who arrive because the price went up.

Then the same crowd unwinds. That is the correction, and it is shorter and messier than the advance, because agreement builds slowly and doubt arrives all at once.

This matters more than it sounds. The wave count is not a pattern imposed on a chart. It is what a crowd changing its mind looks like when you draw it. When a count breaks the rules, it is describing something a crowd could not have done.

The two kinds of wave

Everything in the theory is built from two shapes.

An impulse moves with the trend and takes five waves to do it — three advances with two pull backs between them, numbered one to five.

A correction moves against the trend and takes three, labelled A, B and C.

Five forward, three back. That eight-wave cycle is the whole vocabulary. Every larger structure is these two shapes nested inside each other, and every argument about a chart is an argument about which shape you are looking at.

The distinction sounds simple and is where most of the difficulty lives. A three-wave move against the trend and the first three waves of a new impulse with the trend can look identical while they are forming. Telling them apart is not a matter of pattern recognition. It is a matter of checking which one the rules still permit.

The rules that are not optional

Three rules define an impulse. Not guidelines, not tendencies. If a reading breaks one of them, it is not a weaker reading — it is not an impulse at all, and whatever you were about to conclude from it does not follow.

Wave two never retraces the whole of wave one. If price goes back past where wave one began, wave one was not wave one. The count is dead at that moment, not weakened.

Wave three is never the shortest of one, three and five. It is usually the longest. It is allowed not to be. It is not allowed to be the smallest.

Wave four does not overlap wave one. The territory wave one covered stays covered, with one exception: a diagonal, where the rule is relaxed because the structure is different.

These are worth stating flatly because they are what makes the theory testable. Without them, a wave count is a story about a chart. With them, it is a claim that can be shown to be wrong.

Guidelines are a different thing

Not everything in Elliott Wave analysis is a rule, and treating the two as equivalent is where a lot of bad counting starts.

Alternation is the clearest example. Where wave two is a sharp, deep retracement, wave four tends to be a shallow sideways one — and the other way round. That is a strong tendency and a useful expectation. It is not a law. A count where both corrections take the same shape is unusual, and still legal.

The distinction matters practically. A rule tells you a reading is impossible. A guideline tells you a reading is less likely. Confuse them and you either throw away valid counts or keep invalid ones, and both mistakes feel the same from the inside.

Why Fibonacci turns up

Wave relationships cluster around a handful of ratios — .382, .5, .618, 1.618 and their relatives. Wave two commonly retraces .618 of wave one. Wave three often extends 1.618 of it.

Why this happens is genuinely unsettled. The honest answer is that the proportions recur often enough to be useful and nobody has explained them to everyone's satisfaction. Anyone telling you otherwise is selling something.

What matters is what the ratios are for. They do not identify a count. They grade one. Once the rules have decided which readings are permissible, the Fibonacci relationships tell you which of the survivors is proportioned like the ones that came before — in price, and in how long each wave took.

A count that satisfies the rules but fits no ratio at all is legal and unremarkable. That is worth knowing before you trade it.

Degrees, or why the same chart has more than one answer

A five-wave advance on a daily chart is one wave of a larger structure on the weekly, and contains five smaller waves of its own on the hourly. The pattern repeats at every scale, which is the part people find either elegant or maddening.

Elliott named the scales, from Grand Supercycle down through Cycle, Primary, Intermediate, Minor and on to Subminuette and Micro. The names are a bookkeeping convenience — what matters is that any wave you are looking at is simultaneously part of something larger and made of something smaller.

This is why two competent analysts can look at the same chart and disagree without either being careless. They may be counting at different degrees. It is also why the first question about any wave count is not "is this right" but "at what degree".

Timeframes, and why one chart is not enough

A count that holds on a daily chart and falls apart on the weekly is not a count you should trust. The structures nest, so a reading has to be consistent with the degree above it — a wave three on the daily has to sit somewhere sensible inside whatever the weekly is doing.

Working across two timeframes is not about finding more detail. It is about constraint. Each additional degree removes readings that were legal in isolation, which is the point: the answer you want is the one that survives being checked from more than one direction.

In practice this is why a count on a single chart, however careful, is a weaker claim than the same count nested inside its parent.

Why most wave counts you see are wrong

Because the method rewards motivated reasoning and almost nothing pushes back — not because the people doing it are foolish.

A chart usually permits several legal readings. Picking among them is where judgement enters, and judgement is where a view you already hold gets to put its thumb on the scale. If you are bullish, the count that is bullish looks better. Then wave four overlaps wave one, and rather than discarding the count, it becomes a diagonal. Then the diagonal does not quite work either, so the degree gets adjusted.

Each individual step is defensible. The end result is a count that has been shaped to a conclusion rather than derived from a chart, and it is indistinguishable from a good one in a screenshot.

The tell is what happens when a rule breaks. A count that gets rescued every time it is contradicted was never being tested.

What Waveanism does about it

The problem is not that people cannot apply the rules. It is that applying them consistently, to every possible reading, every time, is more arithmetic than anyone does by hand.

So Waveanism does the arithmetic. It searches the legal readings rather than proposing one, discards anything that breaks a rule before you ever see it, grades what survives against the Fibonacci relationships in price and in time, and tells you the level at which each reading would be invalidated.

It does not tell you what to buy, and it is not built to. When the rules permit nothing, it says so rather than offering the least-bad story.

See the plans How it counts