What is Elliott wave theory?

Elliott wave theory, or the Elliott wave principle, describes price as waves of crowd psychology: optimism pushes a trend forward in five steps, then doubt pulls it back in three. Ralph Nelson Elliott, an American accountant, developed it in the 1930s from stock market charts and published it in 1938 in The Wave Principle.

It faded for decades until A.J. Frost and Robert Prechter published Elliott Wave Principle in 1978, now the standard textbook. Elliott later tied the waves to the Fibonacci sequence: 5, 3 and the full cycle of 8 are all Fibonacci numbers.

The patterns are fractal: every wave splits into smaller waves, and every 5-3 cycle is one wave of a larger pattern. Elliott called these sizes degrees. A five-wave rally on a 5-minute chart can be just wave 1 on the hourly chart, so two traders can label the same move differently.

Impulse wave
A five-wave move in the direction of the larger trend, labeled 1-2-3-4-5. Waves 1, 3 and 5 push with the trend; waves 2 and 4 pull back.
Corrective wave
A move against the larger trend, usually in three waves labeled A-B-C. Common shapes are the zigzag, the flat and the triangle.
Invalidation level
The price that would break one of the three rules. If price trades through it, the current count is wrong and has to be relabeled.
Alternate count
A second labeling that also fits the rules. Elliott analysts keep one in reserve because the preferred count is often wrong.

The 5-3 pattern, wave by wave

The drawing at the top of this page shows the textbook shape: a five-wave impulse up, wave 3 the longest, then an A-B-C correction down. Each wave tends to have a character, listed below: tendencies that help you label a chart, not laws.

A hypothetical count on a stock: wave 1 rallies from $100 to $110, wave 2 pulls back to $104, wave 3 runs to $121, wave 4 dips to $116 and wave 5 tops at $124. Then A falls to $117, B bounces to $121 and C drops to $115. It passes all three rules below: the wave 2 low ($104) holds above $100, wave 3 ($17) is the longest leg, and the wave 4 low ($116) holds above the wave 1 high ($110).

  • Wave 1: the first push off a low, often dismissed as just another bounce.
  • Wave 2: a pullback that gives back much of wave 1, often more than half, never all of it.
  • Wave 3: usually the strongest, broadest leg, and often the longest.
  • Wave 4: a pause, often sideways and shallower than wave 2.
  • Wave 5: the last push to a new high, often on weaker momentum.
  • A, B and C: a first drop, a bounce that can look like the trend resuming, and a final leg down.

The three Elliott wave rules

Elliott analysts separate rules from guidelines. A rule never bends: break one and the count is wrong, however good it looked. The three rules apply to a standard five-wave impulse.

The second rule is often misread: wave 3 need not be the longest, only not the shortest. With wave 1 at $10 and wave 5 at $8, a $9 wave 3 passes; a $7 one does not. The third rule has one known exception: in a less common five-wave pattern called a diagonal, wave 4 usually overlaps wave 1. So an overlap means a diagonal, or wrong labels.

The failure case, hypothetical: you call a rally from $100 to $110 wave 1, and the pullback slides to $99.80. That is more than 100% of wave 1, so it cannot be wave 2; the rally was not a wave 1, or the low was mislabeled. The count dies at one exact price, and that precision is the most useful thing the theory gives a trader.

  1. Wave 2 never retraces more than 100% of wave 1. In the example, the wave 2 low must stay above $100, where wave 1 started.
  2. Wave 3 is never the shortest of the three impulse waves (1, 3 and 5). Measure each from its start to its end.
  3. Wave 4 does not enter the price territory of wave 1 in a standard impulse. In the example, the wave 4 low must stay above $110, the top of wave 1.

Elliott wave guidelines: tendencies, not rules

Guidelines describe what often happens. They help you choose between counts that all pass the rules, and they fail regularly. Treat each one as a reason to look, never as a reason to trade.

In the hypothetical count, wave 2 gave back 60% of wave 1 in a quick drop, while wave 4 drifted sideways and gave back about 29% of wave 3. That is alternation at work. Had both been sharp and deep, the count would still pass the rules, but as a weaker candidate.

  • Alternation: if wave 2 is a sharp, deep pullback, expect wave 4 to be sideways and shallow, and vice versa.
  • Extension: one impulse wave is often much longer than the other two. In stocks, it is most often wave 3.
  • Equality: when wave 3 extends, wave 5 tends to be close in size to wave 1.
  • Momentum: wave 5 often makes its new high on weaker momentum than wave 3, which many traders check with RSI divergence.

How Fibonacci ratios fit Elliott waves

Elliott practitioners measure waves against each other with Fibonacci ratios. The common tendencies: wave 2 often retraces 50% to 61.8% of wave 1, wave 4 often retraces 38.2% of wave 3 or less, and wave 3 often reaches about 1.618 times the length of wave 1, measured from the end of wave 2.

In the hypothetical count, wave 1 is $10. A wave 2 in the 50–61.8% band would end between $105.00 and $103.82; the actual low of $104 sits inside it, the pullback shape in the drawing below. A 1.618 projection for wave 3 is $104 + $16.18 = $120.18, and price reached $121: close, not exact. Wave 4 at 38.2% of wave 3 would be $121 − $6.49 = $114.51; the actual low of $116 was shallower.

Ratios are zones, not exact prices, and they only matter once the count passes the rules: a pullback that pauses at 61.8% and then breaks the start of wave 1 is not a wave 2. The Fibonacci retracement guide shows how to draw the levels on TradingView, and harmonic patterns build on the same ratios.

Turn your phone sideways to read the drawing. Tap to close.

A hand-drawn swing from a low to a high with retracement levels at 23.6%, 38.2%, 50%, 61.8% and 78.6%; the 50–61.8% band is shaded yellow, and a buy signal prints near 61.8%.

Drawing channels for Elliott waves

Elliott noticed that impulse waves often travel inside parallel channels. The common construction: once wave 4 seems to have ended, draw a line connecting the ends of waves 2 and 4, then a parallel line through the end of wave 3. The upper line marks a rough zone where wave 5 may end.

In the hypothetical count, that line runs through the wave 2 low ($104) and the wave 4 low ($116), and its parallel through the $121 top of wave 3 projects a zone for wave 5 above it. Price can stop short of the line or overshoot it in a final burst.

The lower line matters as much: a close back below the 2–4 line after five waves warns that the impulse may be over, or that the count is wrong. The drawing shows the idea: a rising line under three higher lows, a dashed channel line over the highs, and a final candle closing below the line. Its circled numbers count touches, not waves. See trend lines for drawing both lines without forcing them.

Turn your phone sideways to read the drawing. Tap to close.

A hand-drawn uptrend: a rising trendline under three higher lows, with the touches circled 1, 2 and 3, a dashed channel line over the highs, and a final candle that closes below the trendline, labeled break.

Where a count is invalidated, and why that is your stop

Elliott waves rarely tell you where a move will end, but the rules tell you exactly where a count is wrong, and that price is a natural stop. The classic trade joins wave 3 after wave 2 ends: wave 3 is often the strongest leg, and its invalidation level, the start of wave 1, is known in advance.

A hypothetical plan on a 15-minute chart: a stock rallies from $50.00 to $51.50 in five small waves, your wave 1. It pulls back to $50.70, about 53% of wave 1, and a candle then closes back up at $50.95. That close is the confirmation; a pullback that has not turned up is not a trade yet. Entry is $50.95. Wave 2 cannot go below $50.00, so the stop goes just under it, at $49.95: $1.00 of risk per share. A 1.618 projection for wave 3 is $50.70 + $2.43 = $53.13, so the target sits just in front of it, at $52.95: $2.00 of reward, a 1:2 risk/reward ratio.

Size comes from the stop. A $25,000 account risking 0.5% accepts a $125 loss: $125 ÷ $1.00 = 125 shares. The position size calculator does this math for you.

The losing version: price slips under $50.70, then trades through $50.00. The count is dead, the stop fills near $49.95 for −1R, about $125 plus costs, and the rally was probably a three-wave bounce, not a wave 1. Do not move the stop because a new count appears. A tighter stop under the wave 2 low, at $50.65, risks only $0.30, but a deeper wave 2 that is still valid will hit it more often. Gaps can fill a stop beyond its price, so a loss can exceed 1R; see stop loss and take profit.

Turn your phone sideways to read the drawing. Tap to close.

A hand-drawn trade plan: an entry line, a red box down to the stop labeled risk = 1R, a green box up to the target labeled reward = 2R, and a note reading R:R = 1:2.

Why Elliott wave counts are subjective and change

Elliott practitioners say it themselves: at any moment, two or more counts usually fit the rules. The rules eliminate labelings but do not pick one. Where does wave 1 start? Is this pullback wave 2, or wave A of something bigger? Is it a 5-minute wave or part of an hourly one? Each answer moves every label after it.

Counts also change as price unfolds: a wave 5 can extend, a correction can stretch into a triangle, and a clean impulse can turn out, in hindsight, to be part of a correction. Critics argue the rules are loose enough to fit almost any chart after the fact. Treat a count as a hypothesis with an exit, not a forecast.

In practice, keep a preferred count and one alternate, and write down the price that kills each. When the preferred count dies, the alternate suggests what price may be doing instead; it is not a reason to jump back in. Methods like this fail often, so size every trade for being wrong.

How to use Elliott waves on intraday charts

On 5-minute to 1-hour charts, a full count is noisy and any news spike can redraw it. Most traders use Elliott waves for context and keep the trade simple, with the steps below. The drawing shows the location step: price dips into support, rejects it and prints a buy signal on a closed candle.

UpTickFlow, our invite-only TradingView indicator, does not count waves or draw Fibonacci levels. It waits for a candle to close at a level, scores the setup from 0 to 100, shows only A+ setups, draws the entry, stop and target before the trade, and stays quiet in chop. Use it as a second check: if your wave 2 ends where UpTickFlow prints a signal on a closed candle, two independent reads agree. If it stays silent, that is information too.

  1. Read the higher timeframe first: is the hourly chart in an impulse or a correction? See multi-timeframe analysis.
  2. Trade only the clearest moment, usually the end of a wave 2 or of an A-B-C correction in the trend direction.
  3. Demand a location: support, VWAP or the 50–61.8% zone of wave 1. See the support and resistance strategy.
  4. Wait for a candle that closes back in the trend direction. A wick into the zone proves nothing.
  5. Put the stop at the invalidation price and the target before the next level, and size from the stop.
  6. Skip sideways, overlapping chop. If you cannot label it in seconds, there is no trade.

Turn your phone sideways to read the drawing. Tap to close.

A hand-drawn chart with a resistance zone above and a support zone below: price dips into support, rejects it and prints a buy signal.