What Is a Fractal in Trading? How to Use Fractals

Search Fintrev

Search our guides, reviews and glossary.

A fractal in trading is a five-bar price pattern that marks a local turning point. In a bearish fractal, the middle bar has a higher high than the two bars before it and the two bars after it. In a bullish fractal, the middle bar has a lower low than the two bars on each side. The pattern was popularized as an indicator by trader Bill Williams, which is why charting software often lists it as “Williams Fractals.” The name comes from fractal geometry, the branch of mathematics Benoit Mandelbrot developed to describe shapes that repeat at different scales.

A fractal identifies a swing high or swing low after the fact. It cannot appear until two more bars have closed beyond the turning point, so every signal arrives with a built-in delay. Traders use fractals to locate reversal points, place stops, define breakout levels and anchor tools such as Fibonacci retracements. The pattern does not forecast direction by itself.

What Is a Fractal in Trading?

On a price chart, a fractal is the smallest pattern that qualifies as a swing point. If the third of five consecutive bars sticks out above its neighbors, it is an up fractal, also called a bearish fractal or fractal high. If it dips below its neighbors, it is a down fractal, also called a bullish fractal or fractal low.

The terms can be confusing because platforms name them differently:

The logic behind the pattern is about balance between buyers and sellers. At a fractal high, buyers pushed price to a peak and then failed to exceed it for two full bars. At a fractal low, sellers drove price to a trough and then failed to break it for two bars. In both cases the market tested an extreme and backed away from it.

A fractal only looks at highs and lows. Opens, closes, candle colors and volume play no part in the definition. A fractal high can form on a bar that closed lower than it opened, and a fractal low can form on a strong green candle. That separates fractals from candlestick reversal patterns, which depend on the shape and color of the bodies.

How a Fractal in Trading Works

The five-bar rule

The standard definition has four components:

1. A central bar. This bar holds the extreme: the highest high for a bearish fractal, the lowest low for a bullish fractal.
2. Two bars to the left. Their highs (bearish case) or lows (bullish case) must be less extreme than the central bar.
3. Two bars to the right. The same condition applies to the two bars that follow.
4. Confirmation on the close of the fifth bar. Until the second bar to the right closes, a new high or low can still cancel the pattern.

The neighboring bars do not have to form a neat staircase. In the bearish case, the two left bars do not need rising highs, and the two right bars do not need falling highs. All four only need highs below the central high. The symmetrical textbook drawing is one possible arrangement among many valid ones.

A single bar can be both an up fractal and a down fractal. This happens when a wide-range bar has both the highest high and the lowest low of its five-bar window, often during a volatile session that swept liquidity on both sides. Such a bar marks indecision rather than a clean turning point.

Equal highs, equal lows and longer windows

Real price data often contains ties. If the central bar’s high equals the high of a neighbor, the strict definition is not met. Implementations differ: some indicators require strictly higher highs, some allow ties on one side, and some extend the window until the tie is resolved. That is why fractals are sometimes described as patterns of “five or more” bars. If two charting platforms show slightly different fractal markers on the same data, a different tie rule is the likely cause.

Many indicators also let you change the number of bars on each side. The default in Williams’ version is two. Raising it to three, four or more bars per side produces fewer, more significant swing points, because the central bar must dominate a wider window. The cost is a longer delay. A fractal with five bars on each side cannot be confirmed until five bars after the turning point.

Why fractals lag, and what “repainting” means

Lag is part of the definition. Say a bearish fractal’s central bar closes on Monday on a daily chart. Tuesday and Wednesday must both print lower highs before the pattern exists. The signal is known only at Wednesday’s close, yet the marker is drawn on Monday’s bar.

Looking back, every fractal arrow seems to sit right at the top or bottom of a swing, which makes the tool look more precise than it was. In real time, a trader saw the arrow two bars late, after price had moved away from the extreme. Some scripts also draw a tentative marker before the fifth bar closes and remove it if a new extreme appears. In any backtest or chart review, count where price was when the fifth bar closed, not where the arrow is.

Interpreting a fractal in trading

The arrow convention causes frequent mistakes. A bearish fractal is usually drawn with an arrow pointing up, sitting above the high. A bullish fractal gets an arrow pointing down, below the low. The arrow points to the extreme, not to the expected direction. In an uptrend where you want long entries from pullback lows, you look for the down arrows. In a downtrend where you want short entries from bounce highs, you look for the up arrows.

There are two main readings of the same fractal, and they lead to opposite trades.

• Reversal reading. A bearish fractal suggests a possible top, so a trader may sell short or exit a long position. A bullish fractal suggests a possible bottom, so a trader may buy. Stops go beyond the fractal’s extreme.
• Breakout reading. This is closer to how Williams used fractals. The most recent fractal high is a level where buyers failed. If price later trades above it, the failure has been overcome and the move may continue. A buy stop placed a tick above the last up fractal, or a sell stop below the last down fractal, turns each fractal into a breakout trigger.

Neither reading is correct in every market. The reversal reading tends to fit ranges and pullbacks within a trend. The breakout reading fits trending conditions where old swing points keep giving way.

Why Is a Fractal in Trading Important?

Fractals offer an objective, rule-based way to mark swing highs and lows. Two traders looking at the same chart with the same settings will mark the same points, which is not true when swing points are picked by eye. That consistency supports several common tasks.

Support and resistance. Fractal highs and lows mark levels where price recently turned. A cluster of fractal lows near the same price is a support zone defined by repeated rejection, not by a line drawn freehand.

Stop placement. A long trade’s stop is often placed below the most recent bullish fractal, and a short trade’s stop above the most recent bearish fractal. The logic is structural: if price moves beyond the swing point that defined the trade, the setup has failed.

Trailing stops. In a trend, a trader can move the stop to each new fractal low (for longs) or fractal high (for shorts) as it confirms. The stop then follows the market’s swing structure instead of a fixed distance.

Market structure. A sequence of higher fractal highs and higher fractal lows describes an uptrend in mechanical terms. Lower fractal highs and lower fractal lows describe a downtrend. When a new fractal low forms below the previous one in an uptrend, the structure has changed. Traders who use market-structure methods (including those under the “smart money concepts” label) often use fractals or similar swing-point rules to define these highs and lows.

Anchoring other tools. Fibonacci retracements, trendlines and channels all need start and end points. Using confirmed fractals as anchors removes some of the guesswork about which high or low to choose.

Fractals can be applied to any instrument with price bars: currency pairs, stocks, indices, commodities and crypto. They also work on any time frame, from one-minute charts to monthly charts. A fractal on a weekly chart reflects a turning point that held for two weeks on each side. A fractal on a five-minute chart reflects a pause of ten minutes. The pattern is identical, but the significance is not.

Example of a Fractal in Trading

The following is a hypothetical example for illustration only. The prices are invented and do not represent real market data.

Five consecutive daily bars on a currency pair:

Step 1: check the highs. Bar 3’s high (1.0925) is above the highs of bars 1 and 2 (1.0850 and 1.0880) and above the highs of bars 4 and 5 (1.0895 and 1.0870). Bar 3 is a bearish fractal, and its marker goes above bar 3.

Step 2: check the lows. Bar 3’s low (1.0840) is not the lowest of the five. Bar 1’s low is 1.0790. So bar 3 is not also a bullish fractal.

Step 3: note when the signal exists. The fractal is confirmed at the close of bar 5, at 1.0810. By then, price is 115 pips below the fractal high. A trader using the reversal reading who sells at 1.0810 has already missed that part of the decline.

Step 4: work out the risk. If the stop goes 5 pips above the fractal high, at 1.0930, the distance from entry is 120 pips. Assume a $10,000 account, a 1% risk limit ($100), and a pip value of $10 per standard lot (which applies to a pair quoted in US dollars, such as EUR/USD, at 100,000 units per lot). The position size is $100 ÷ (120 pips × $10) ≈ 0.083 lots, or about 0.08 lots. A position size calculator does this arithmetic for other pairs and account currencies.

A reversal entry on a fractal comes late, and the stop beyond the extreme can be wide. That is why many traders take the signal only with supporting evidence, or switch to the breakout reading. In this example, a sell stop under the most recent bullish fractal below price would trigger only if the market broke that swing low.

How to Use Fractals in Trading

Reading fractals in context

A fractal carries more weight in some locations than others. A bullish fractal at a known support level, or at the end of a pullback within an established uptrend, has a clear reason to hold. One that forms in the middle of a strong downtrend, far from any support, is more likely a brief pause before the next leg lower. The same rule applies in reverse: a bearish fractal at established resistance, or after a bounce inside a longer downtrend, aligns with the larger picture, while one during a strong uptrend often just marks the top of a minor pullback.

Reducing the number of signals

On an intraday chart, new fractal markers appear constantly, most of them during sideways chop. Common ways to reduce this noise include:

• Move to a higher time frame. Fractals on daily or weekly charts appear less often and reflect turning points that held for longer.
• Increase the bars per side. Requiring three or more bars on each side of the central bar filters out minor swings.
• Trade only with the trend. In an uptrend, take only bullish fractals as entries and use bearish fractals for exits or trailing stops.
• Use multiple time frames. Identify the trend and key fractal levels on a higher time frame, then use lower time frame fractals for timing entries near those levels.

Combining fractals with Fibonacci retracements

Fibonacci retracements show several possible levels, typically 23.6%, 38.2%, 50% and 61.8% of a prior move, but not which one will hold. A fractal alone does not show which turning point matters. Combined, each narrows the other: the trader takes a setup only when a fractal forms at or near a retracement level.

Hypothetical example. A pair rallies from a confirmed fractal low at 1.0500 to a confirmed fractal high at 1.1000, a 500-pip move. The 61.8% retracement sits at 1.1000 − (0.618 × 0.0500) = 1.0691. Instead of buying as soon as price touches 1.0691, the trader waits. Price drops to 1.0688, and two bars later a bullish fractal confirms with its low at 1.0688. The fractal shows that sellers failed to extend the decline near the Fibonacci level. The trader buys, places the stop below 1.0688, and plans to exit or tighten the stop when a bearish fractal suggests the rally is stalling.

Using confirmed fractals as the retracement anchors keeps the method consistent. The Fibonacci retracement guide covers how the levels are drawn.

Using fractals with the Alligator

Bill Williams designed fractals to work alongside his Alligator indicator. It consists of three smoothed moving averages, each shifted forward on the chart:

• Jaw (often blue): 13-period smoothed moving average, shifted 8 bars forward
• Teeth (often red): 8-period smoothed moving average, shifted 5 bars forward
• Lips (often green): 5-period smoothed moving average, shifted 3 bars forward

When the three lines are intertwined, the market is ranging (the “sleeping” alligator), and fractal signals are more likely to fail. When the lines fan out in order, the market is trending. In an uptrend, the lips are above the teeth and the teeth above the jaw. Williams’ rule was to take a fractal breakout only on the trend side of the teeth: a buy above an up fractal counts only if that fractal sits above the red line, and a sell below a down fractal counts only if it sits below it. The Alligator acts as a trend filter and the fractal supplies the trigger.

Other indicators used with fractals

• Moving averages. A 50-period or 200-period moving average defines trend direction, and fractals mark pullbacks into or bounces from it. An exponential moving average reacts faster than a simple one.
• Momentum oscillators. A bullish fractal that forms while the Relative Strength Index is oversold, or shows bullish divergence, combines a structural low with weakening downside momentum. MACD can play a similar role.
• Volume. Rising volume on a break of a fractal level suggests real participation behind the move. Spot forex has no centralized volume, so traders substitute tick volume or futures volume.
• Volatility measures. A stop exactly at the fractal extreme can be hit by ordinary noise. A buffer based on the Average True Range ties the stop distance to how much the instrument normally moves.
• Price action. A fractal low that coincides with a hammer candle at a long-tested support level has more supporting evidence than one that forms in open space.

Limitations of a Fractal in Trading

Lag. A fractal confirms only two bars after the extreme, so entries come late and stops beyond the extreme are wider than the marker suggests. Larger bar settings make the lag worse.

False signals. Every short pause creates a fractal. In a ranging or choppy market, fractal highs and lows alternate rapidly, and many trigger trades that reverse at once. The pattern says only that a bar was more extreme than its four neighbors, not that direction has changed for good.

Weak as a standalone tool. Because the definition ignores trend, location, momentum and volume, most fractal methods rely on another tool for context. The fractal answers “where is the recent swing point?” and leaves “should I trade it?” to the rest of the analysis.

Fractals vs. Related Concepts

Fractal vs. swing high or swing low. A swing high is any local peak and a swing low any local trough. A fractal is a swing point defined by a fixed rule (two bars on each side by default). Every fractal is a swing point, but a swing point that a trader picks by eye may not meet the fractal rule.

Fractal vs. pivot points. Classic pivot points are levels calculated from the prior period’s high, low and close, and they are known before the new period begins. Fractals are discovered from price action after it happens. Some platforms use “pivot high” and “pivot low” for what is essentially a fractal with adjustable bars per side, so the term depends on the software.

Fractal vs. candlestick reversal patterns. Patterns such as hammers, engulfing candles and dojis depend on the open, close and body shape. Fractals depend only on highs and lows across five bars.

Fractal indicator vs. fractal geometry. The five-bar indicator is a simple pattern rule that borrows its name from a mathematical field about self-similar structures. It does not calculate any fractal dimension. Quantitative work on the fractal nature of markets, such as estimating the Hurst exponent or modeling prices with fractional Brownian motion, is a separate discipline.

Fractal Theory in Finance

Mandelbrot and self-similarity

Benoit Mandelbrot introduced fractals in the 1970s as structures that show similar patterns at different scales. He argued that financial prices behave this way: hourly, daily and weekly charts look alike, so an observer cannot easily tell the time scale from the shape. He also criticized standard models, including modern portfolio theory, for assuming normally distributed price changes, which underestimates how often extreme moves occur. His work drew renewed attention after the 2007 to 2008 financial crisis.

Mandelbrot wrote that fractal models do not forecast a rise or drop on a specific day. They describe the probability of different kinds of market behavior and help prepare for large, abrupt changes. The five-bar indicator likewise describes structure, not certainty.

Williams built his trading approach on chaos theory, which studies systems that follow deterministic rules yet behave in ways that look random. Price action looks chaotic at any single moment, while swings, pauses and reversals recur at every scale. The view is reflected in the title of his 1998 book, New Trading Dimensions: How to Profit from Chaos in Stocks, Bonds, and Commodities.

The rationale: market psychology

Supporters argue that traders react to similar conditions in similar ways, through fear, greed, herding and memory of past price levels, and that these reactions create repeating structures. Fractal levels then become partly self-reinforcing, because many participants see the same swing points and place orders around them.

The Fractal Market Hypothesis

The Fractal Market Hypothesis (FMH), developed by Edgar Peters in the 1990s, is an alternative to the Efficient Market Hypothesis, which holds that prices move randomly in response to new information. The FMH links market stability to investors with many different time horizons taking the other side of each other’s trades. When they start acting alike, for example when long-term holders trade like short-term ones during a panic, liquidity dries up and markets become unstable.

The FMH is a theory about market structure and stability. It does not prescribe trades and does not validate any particular chart indicator.

Can Fractals Predict Future Price Movements?

No. A fractal describes where price turned, confirmed after the turn. It carries no information about how far the next move will go or whether the swing will hold. At best it is a probability tool whose usefulness depends on the trend, location, time frame and confirmation around it.

FAQs

The Bottom Line

Fractals suit traders who want a consistent, rule-based way to mark swing points for stops, trailing stops, breakout levels and Fibonacci anchors. Because signals lag and multiply in choppy markets, the pattern works best alongside trend, location and momentum context. A fractal shows where price recently turned and cannot predict how far the next move will go.