Elliott waves and harmonic patterns: principles and limits

Some approaches to technical analysis go further than classic chart patterns. They look for a hidden order in market moves, a kind of rhythm that supposedly repeats at every scale. The two best known are Elliott wave theory and harmonic patterns. They have passionate followers and fierce critics. This lesson gives you the basics, then takes an honest look at their limits.

The principle: 5 waves, then 3

In the 1930s, Ralph Nelson Elliott proposed that markets move in cycles driven by crowd psychology. In the direction of the main trend, price would advance in 5 waves: three impulse waves (1, 3 and 5) separated by two corrective waves (2 and 4). Then comes a larger correction in 3 waves, labelled A, B and C.

• Wave 1: the start of the move, often quiet. Few people believe in it.
• Wave 2: a correction that sows doubt. It can be deep.
• Wave 3: usually the strongest and longest. The wider public joins in.
• Wave 4: a pause, often sideways and slow.
• Wave 5: the final push, sometimes with fading momentum (an RSI divergence, for example).
• Waves A, B, C: the correction of the whole move, with a false rally (B) in the middle.

The pattern is said to be fractal: each wave breaks down into smaller waves, and each complete cycle is just one wave of a bigger cycle. That's both the appeal of the theory and its main difficulty.

The three rules that aren't up for debate

The theory has lots of flexible "guidelines", but only three strict rules for a standard impulse wave. If any of them is broken, the count is wrong and you need to find another one.

1. Wave 2 never retraces beyond the start of wave 1. If price breaks that point, there was no wave 1.
2. Wave 3 is never the shortest of the three impulse waves 1, 3 and 5.
3. Wave 4 doesn't enter the price territory of wave 1 (except in special cases called diagonals).

Analysts often pair waves with Fibonacci ratios. Wave 2 often retraces around 50% to 61.8% of wave 1. Wave 3 is often projected at 161.8% of wave 1. These aren't rules, just observed reference points, with plenty of exceptions.

Harmonic patterns in brief

Harmonic patterns take the Fibonacci logic even further. They describe five-point structures, labelled X, A, B, C and D, where each leg has to respect precise ratios relative to the others. The best-known names are the Gartley, the Bat, the Butterfly and the Crab.

The principle is always the same: point D is a potential reversal zone, calculated from several ratios that converge. In the Gartley, for example, point B sits around 61.8% of the XA leg and point D around 78.6% of that same leg. The trader waits for price at D, looks for a rejection signal, and places the stop just beyond it.

The limits you absolutely need to know

Here's why these methods are so divisive. The first limit is subjectivity. Where does wave 1 start? Is this zigzag a wave 4 or the start of an A correction? Two serious analysts can come up with opposite counts on the same chart, and both will respect the rules. Harmonic patterns, for their part, allow a tolerance on each ratio, which leaves a lot of room for interpretation.

• Endless recounting: when price contradicts the scenario, it's tempting to recount until the chart fits. A theory that explains everything after the fact predicts nothing.
• Hindsight bias: on a past chart, the waves look obvious. In real time, you never know which wave you're in.
• No solid proof: there's no rigorous, widely accepted demonstration that these structures give a measurable edge once costs are taken out.
• Complexity: the more special cases a method has, the easier it is to use it to justify a decision you've already made.

Used with humility, these approaches have some value: they force you to think in scenarios, with a clear invalidation level. Used as a crystal ball, they lead to overconfidence. A wave or a pattern is only a hypothesis, one probability among others, never a certainty. Leveraged trading loses money for most retail traders, and no theory changes that without risk management.