Elliott Wave, A Reality Check

[Fairman]

Elliott Wave, A Reality Check

Elliott Wave theory proposes that markets move in repeating patterns: five waves in the direction of the trend, followed by three corrective waves. It's popular, and also controversial.

diagram.png (37.03 KiB) Viewed 50 times

The basic idea:

- Impulse waves (1, 2, 3, 4, 5): waves 1, 3, and 5 move with the trend, and waves 2 and 4 retrace.
- Corrective waves (A, B, C): a three-wave pullback against the trend.

Key rules traders cite:

- Wave 2 doesn't retrace beyond the start of wave 1
- Wave 3 is never the shortest
- Wave 4 doesn't overlap wave 1's territory (in standard impulses)

Why it's debated:

- Wave counts are subjective, and different analysts count the same chart differently
- You can almost always find a count that fits after the fact
- Fixed rules can feel flexible when exceptions are allowed

How to use it, if you choose to:

1. Treat it as a framework, not a prediction machine
2. Use clear invalidation points
3. Combine with structure, levels, and risk management
4. Test your counts on past charts honestly

Many traders find its main value is in recognizing that trends move in phases with pullbacks, rather than in exact wave counts.

If a method can't be tested clearly, be careful about risking real money on it.

[LondonNewsTrader]

Fair write-up, and the subjectivity point is the real issue. Two people can agree on every rule you listed and still disagree on where wave 1 started, which means they disagree on everything after it.

What I'd add is that a lot of the clean "wave 3" moves people show on charts are really event moves. A CPI surprise or a surprise rate decision produces exactly the long, impulsive leg that the theory says should be the strongest wave, and then the count gets built around it after the fact. If you mark the release times on those charts, the waves often line up with the calendar better than with any Fibonacci ratio.

If someone wants to use it, your point 2 is the one that keeps it honest. Write down the invalidation level before the trade, which in practice is usually the start of wave 1, and accept the loss when it breaks rather than relabelling the count.

[FXS Prop Desk]

A fair reality check. A practical way to use Elliott Wave with less ambiguity is to focus on its hard rules, e.g. wave 2 not retracing past the start of wave 1, and wave 4 not entering wave 1's territory in an impulse, and to use them as stop levels. If the rule breaks, the count is wrong and the trade is closed, with no recounting.

That turns a subjective tool into one with clear invalidation. Traders who use it this way often treat Elliott only as a map and let price structure decide entries.

[Shadow Trader]

Using the hard rules as stops is the cleanest version, but on a scalping timeframe the distance can be awkward. Buying the end of wave 2 on M15 with the stop under the start of wave 1 can easily mean 30 pips or more, which is swing risk on scalping size.

The wave 4 entry is where the rules actually help a scalper. If wave 1 topped at 1.1641 and the wave 4 low holds at 1.1655, the overlap rule puts invalidation at 1.1641, so the stop is about 15 pips with a buffer and a clear reason behind it.

For wave 5 I aim modestly, usually for a move equal to wave 1, rather than any extension. Fifth waves in FX often stall around a release or a session close, so expecting a big one is how the profit leaks back.

LondonNewsTrader's point about event legs matches what I see. When I mark the calendar on old counts, the move labelled wave 3 starts within an hour of a release more often than not.

My personal rule is that a count relabelled twice in one session is dropped for the day. Does anyone else use a limit like that, or do you just keep recounting?