What Elliott Wave is
The Elliott Wave Principle is a theory of how market trends unfold. It was developed by Ralph Nelson Elliott in the 1930s. He argued that crowd psychology moves in repeating patterns that can be seen at every scale, from minutes to decades.
The basic structure is 5-3:
- Five waves in the direction of the larger trend, labeled 1, 2, 3, 4, 5. This is an impulse. Waves 1, 3 and 5 move with the trend; waves 2 and 4 are pullbacks.
- Three waves against it, labeled A, B, C. This is a correction.
A full cycle has eight waves. Each wave is itself made of smaller waves, and each completed cycle becomes one wave of a larger pattern. This property is called fractal.
Why traders use it:
- Context. It suggests where you might be in a trend: early, middle or late.
- Scenarios. It turns a chart into “if this, then that” plans.
- Invalidation. Its hard rules tell you precisely where a count, and therefore a trade idea, is wrong.
That third point is the most valuable. Elliott Wave is weak at predicting and strong at defining risk.
Rules vs guidelines
Beginners mix these up. The distinction is the foundation of using Elliott properly.
The three hard rules of an impulse
| Rule | What it means | If broken |
|---|---|---|
| Wave 2 never retraces more than 100% of wave 1 | Wave 2 cannot go beyond the start of wave 1 | The count is wrong |
| Wave 3 is never the shortest of waves 1, 3 and 5 | Wave 3 can be shorter than one of them, but not both | The count is wrong |
| Wave 4 does not overlap wave 1’s price territory | In an impulse, the wave 4 low (in an uptrend) stays above the wave 1 high | The count is wrong, or the pattern is a diagonal |
Diagonals are a special five-wave pattern, usually found in wave 5 or wave C (ending diagonals) or wave 1 or A (leading diagonals), where wave 4 can overlap wave 1. They look like a wedge. Use them sparingly; calling everything a diagonal is a common way to rescue a broken count.
Guidelines (tendencies, not laws)
- Alternation: if wave 2 is sharp and deep, wave 4 tends to be sideways and shallow, and vice versa.
- Extension: usually one of waves 1, 3 or 5 is extended. On many markets it is wave 3. Some traders argue commodities, including gold, show extended fifth waves more often than stock indices, but no verified figure exists for this.
- Channeling: an impulse often fits a channel drawn across the ends of waves 2 and 4.
- Wave 4 often ends in the territory of the previous wave 4 of one lower degree.
- Equality: when wave 3 is extended, waves 1 and 5 tend to be similar in size.
Break a guideline and the count may still be fine. Break a rule and it is not.
How it behaves on gold (XAUUSD)
Gold has features that both help and complicate wave counting:
- Macro-driven trends. Gold’s large trends are often driven by real yields, the US dollar and central bank buying. These trends can produce clean, extended impulses on daily and weekly charts.
- News spikes. A CPI or payrolls release at 8:30 New York can create a single candle that spans what should be several sub-waves. Intraday counts around news are often unreliable.
- Stop runs. Gold frequently pokes beyond an obvious high or low before reversing. This can look like a truncated fifth wave, a failed wave, or a flat’s B wave. It makes short-term counts messy.
- Round numbers. Waves often end near round levels such as 4,000, 4,100 or 4,200, partly because many orders sit there.
- Session rhythm. On intraday charts, Asian sessions often contain corrective (three-wave or triangle) structures, while London and New York sessions carry impulsive legs.
Elliott practitioners often say commodity fifth waves can be powerful and emotional, especially during crisis-driven rallies. Treat that as an observation to test on gold, not as a rule.
How to count waves step by step
- Start on a higher timeframe. Weekly or daily. Identify the dominant trend and the most obvious completed swings.
- Find an impulse. Look for a move with strong, overlapping-free legs. Mark its start.
- Label the largest waves first. Mark the obvious 1-2-3-4-5 or A-B-C at the highest degree you can see.
- Check the three rules. If any rule fails, the count is wrong. Try another.
- Check guidelines. Does wave 3 look strongest? Is there alternation between 2 and 4? Does it fit a channel?
- Drop a timeframe to confirm internal structure. An impulse wave should subdivide into five; a zigzag’s A and C waves into five each.
- Write a primary count and an alternate count.
- Mark invalidation levels for each count.
- Project targets with Fibonacci ratios as zones, not points.
Corrective patterns
| Pattern | Structure | Shape | Notes |
|---|---|---|---|
| Zigzag | 5-3-5 (A-B-C) | Sharp | B retraces part of A; C often equals A or extends to 161.8% of A |
| Flat | 3-3-5 | Sideways | B retraces most or all of A (often 90% or more); C ends near or beyond A’s end |
| Expanded flat | 3-3-5 | Sideways with spikes | B goes beyond A’s start; C goes well beyond A’s end. Very common in stop-run markets like gold |
| Triangle | 3-3-3-3-3 (A-B-C-D-E) | Contracting | Usually appears in wave 4 or wave B; the next move is often a sharp thrust |
| Combinations | W-X-Y (two or more patterns joined) | Sideways and long | A frequent source of confusion and recounting |
Fibonacci relationships between waves
These are the ratios traders most often look for. They are guidelines, not probabilities.
| Wave | Common relationship |
|---|---|
| Wave 2 | Retraces 50% to 61.8% of wave 1 (sometimes 78.6%) |
| Wave 3 | Extends to 161.8% of wave 1 (sometimes more) |
| Wave 4 | Retraces 23.6% to 38.2% of wave 3 |
| Wave 5 | Equals wave 1, or 61.8% of the distance from the start of wave 1 to the end of wave 3 |
| Wave C | Equals wave A, or 127.2% to 161.8% of wave A |
| Triangle legs | Each leg often around 61.8% of the previous one |
For the mechanics of drawing retracements and extensions, see our Fibonacci guide.
Variations and settings: from beginner to pro
Beginner
- Learn to recognize impulse vs correction on daily and H4 charts.
- Memorize the three rules and apply them strictly.
- Do not trade counts yet. Write counts in a journal and see how they evolve.
Intermediate
- Use wave 2 and wave 4 pullbacks as entry areas with rule-based invalidation.
- Keep a primary and alternate count for every chart you analyze.
- Combine wave counts with RSI divergence: momentum divergence between waves 3 and 5 is a classic observation.
Advanced and pro
- Multiple degrees. Track how an H1 count fits inside an H4 wave, which fits inside a daily wave.
- Recognize expanded flats and combinations quickly, since they are common in gold’s stop-run behavior.
- Use time as well as price. Some traders compare the duration of waves 2 and 4 as part of alternation.
- Probability thinking. Assign rough confidence to each count and size positions accordingly.
3 setups traders use
Educational templates only. Each relies on a count you could be wrong about, so the invalidation is the core of the setup.
Setup 1: Wave 3 entry from a wave 2 pullback
| Element | Rule |
|---|---|
| Context | A clear, five-wave first leg (wave 1) away from a significant low or high, ideally after a structure break. |
| Trigger | A three-wave pullback into the 50 to 61.8% retracement of wave 1, followed by a lower-timeframe break of structure in the trend direction. |
| Entry | On the structure break, or on a retest of the break level. |
| Stop (invalidation) | Beyond the start of wave 1. This is the rule-based invalidation: wave 2 cannot retrace more than 100%. |
| Targets | 127.2% then 161.8% extension of wave 1, projected from the end of wave 2. |
| Management | Move stop to break-even once price clears the wave 1 extreme. |
Setup 2: Wave 5 entry from a wave 4 pullback
| Element | Rule |
|---|---|
| Context | A clear, extended wave 3. Wave 4 is forming, ideally alternating with wave 2. |
| Trigger | Wave 4 reaches the 23.6 to 38.2% retracement of wave 3 and shows a lower-timeframe reversal. |
| Entry | On the reversal confirmation. |
| Stop (invalidation) | Beyond the wave 1 extreme (wave 4 cannot overlap wave 1). Many traders use a tighter stop beyond the wave 4 low. |
| Targets | Wave 5 = wave 1 length, projected from the end of wave 4. |
| Management | Fifth waves can truncate. Take profits early and watch for divergence. |
Setup 3: Wave C completion (counter-trend or trend resumption)
| Element | Rule |
|---|---|
| Context | A three-wave correction against the larger trend, with a clear A and B. |
| Trigger | Wave C reaches 100% of wave A (or 127.2% to 161.8%), often at a support zone, and a lower-timeframe impulse forms back in the trend direction. |
| Entry | On the first pullback after that impulse. |
| Stop (invalidation) | Beyond the end of wave C. |
| Targets | The start of wave A, then the prior trend extreme. |
| Management | If the correction extends into a W-X-Y combination, exit and recount. |
Elliott Wave counts are hypotheses. On gold, counts on M15 and lower can change several times in one session, especially around 8:30 New York data and FOMC decisions. Never widen a stop because “it might be an expanded flat”. When the invalidation level breaks, the trade is wrong. Exit first, recount later.
Worked example (hypothetical)
Account: $15,000. Risk per trade: 1% ($150).
On a hypothetical H1 chart, gold bottoms at 4,149.2 after a decline and rallies in five small waves to 4,176.0. This is labeled as a potential wave 1 (length $26.80).
The 50% retracement of wave 1 is 4,162.6 and the 61.8% retracement is 4,159.4. Price pulls back in three waves to 4,161.0, then breaks a minor M15 lower high at 4,166.5.
- Entry: 4,166.0 on the retest of the broken M15 high.
- Stop: 4,148.0, below the start of wave 1 (4,149.2) with a buffer. Risk = $18.00 per ounce.
- Position size: $150 ÷ $18 = 8.3 ounces. Round down to 0.08 lots ($8 per $1 move). Actual risk = $144.
- Target 1 (127.2% extension of wave 1 from the wave 2 low): 4,161.0 + (26.8 × 1.272) = about 4,195.1. Gain about $29 per ounce = $232.
- Target 2 (161.8% extension): 4,161.0 + (26.8 × 1.618) = about 4,204.4.
- Management: move the stop to 4,166 once price breaks above 4,176 (the wave 1 high). Take half at Target 1.
Note the wide stop. Using the true rule-based invalidation forces a small position. Some traders use a tighter stop below the wave 2 low (4,161) with a larger size, accepting that they may be stopped out even if the count is right.
Combining it with other concepts
- Fibonacci: the main tool for projecting wave relationships.
- Market structure: a break of structure after wave 2 helps confirm that wave 3 may be starting.
- Multi-timeframe analysis: essential for tracking wave degrees.
- RSI divergence and MACD: momentum often peaks in wave 3, with divergence into wave 5.
- Trendlines and channels: Elliott channels help anticipate wave 4 and wave 5 endings.
- Support and resistance: wave endings that line up with known zones carry more weight.
Best timeframes and sessions
| Timeframe | Use |
|---|---|
| Weekly and daily | Main trend count; most reliable structure |
| H4 | Swing trading waves 3 and 5 |
| H1 | Entry timing; internal subdivision |
| M15 and lower | Fine-tuning only; very noisy on gold |
Sessions: impulsive legs on intraday charts often start at the London open (about 03:00 New York / 07:00 to 08:00 GMT) or the New York morning (about 08:30 to 11:00 New York / 12:30 to 15:00 GMT). Corrective triangles and flats often form during the Asian session. Swing traders using daily and H4 counts can mostly ignore session timing, except for news risk. See our swing trading guide.
Common mistakes
- Breaking the rules to keep a count. Fix: when a rule breaks, the count is wrong. Recount.
- Labeling everything a diagonal or a combination. Fix: use these labels only when the internal structure clearly supports them.
- Counting only one scenario. Fix: always write a primary and alternate count.
- Counting on M1 or M5 first. Fix: start on the daily and work down.
- Treating Fibonacci targets as certain. Fix: use them as zones and manage the trade with structure.
- Ignoring news. Fix: do not rely on intraday counts through 8:30 New York data or FOMC releases.
Limitations and honest caveats
- High subjectivity. Experienced analysts often disagree on the same chart. The method allows many valid counts.
- Hindsight is easy. Past charts look perfectly labeled; the right edge of the chart never does.
- No verified predictive edge. There is no reliable public evidence that Elliott Wave predicts gold better than other methods. Its value is in planning and risk definition.
- Guidelines are not statistics. “Wave 3 often reaches 161.8%” is a commonly cited tendency, not a measured probability for gold.
- Counts change. Be prepared to change your view often. If that feels uncomfortable, Elliott Wave may not suit your style.
Pro tips
- Trade the invalidation, not the label. The value of a count is the precise point where it fails.
- Wave 3 has a signature. It usually shows the strongest momentum, gaps or fair value gaps, and breaks obvious structure. If the “wave 3” is weak, reconsider.
- Expanded flats are gold’s favorite trap. A B wave that sweeps the prior high, followed by a sharp C wave, is a common stop-run pattern.
- Use alternation to plan wave 4. If wave 2 was a sharp zigzag, prepare for a sideways wave 4 and do not chase the first dip.
- Fifth waves can truncate. Take partial profits early in wave 5, especially when momentum diverges.
- Keep a count journal. Record your primary and alternate counts with dates. Reviewing how often your primary was right is humbling and useful.
Quick checklist
- Have I started my count on the daily or weekly chart?
- Does my count obey all three impulse rules?
- Do I have a written alternate count?
- Does the internal structure (5 or 3) match my labels?
- Is my entry in a wave 2, wave 4 or end-of-C area?
- Is my stop at or beyond the rule-based invalidation?
- Are my Fibonacci targets zones, not exact prices?
- Is my position size calculated from the stop distance?
- Is there high-impact news that could distort the structure?
- Will I exit, not rationalize, if the invalidation level breaks?
Frequently asked questions
What are the three rules of Elliott Wave?
In an impulse, wave 2 never retraces more than 100% of wave 1, wave 3 is never the shortest of waves 1, 3 and 5, and wave 4 does not overlap the price territory of wave 1. If any rule is broken, the count is wrong and must be revised.
Does Elliott Wave work on gold?
Gold forms clear impulsive and corrective swings, so many traders apply Elliott Wave to it. However, the method is subjective, counts often change, and there is no verified evidence that it predicts gold better than other approaches. It works best as a framework for planning scenarios and invalidation.
Which Elliott wave is the strongest?
Wave 3 is often the longest and strongest wave in an impulse, frequently extending to around 161.8% of wave 1. That is a guideline, not a rule; the only rule is that wave 3 cannot be the shortest impulse wave.
What is the difference between a zigzag and a flat?
A zigzag is a sharp 5-3-5 correction where wave B retraces only part of wave A. A flat is a sideways 3-3-5 correction where wave B retraces most or all of wave A, and wave C ends near or beyond the end of wave A.
What Fibonacci levels are used in Elliott Wave?
Common guidelines include wave 2 retracing 50 to 61.8% of wave 1, wave 3 extending to 161.8% of wave 1, wave 4 retracing 23.6 to 38.2% of wave 3, and wave 5 equaling wave 1. These are tendencies people look for, not fixed outcomes.
Why do Elliott Wave counts keep changing?
Because the same price action can often be counted in more than one valid way, and new bars can invalidate the preferred count. Experienced practitioners keep a primary and an alternate count and switch when a rule is broken.
Sources and further reading
- CME Group — Gold futures overview — Contract specifications and trading hours for COMEX gold
- CFTC — Commitments of Traders — Weekly futures positioning, including gold
- Investor.gov — Stop order — How stop orders work, and their limits in fast markets
External sites are independent of XAUUSD Global. Check dates and data yourself.