Candlestick Patterns Explained: Types, Examples and What the Research Says

Illustration of candlestick chart patterns fading into dots.

A hammer appears after a price decline. A few candles later, the market rises. Look back at the chart, and the explanation seems obvious: the pattern signaled a reversal. But look at another chart and you may find the same formation followed by a further decline—or no meaningful move at all. That raises a question that a list of pattern names cannot answer: does recognizing a candlestick pattern help you anticipate what happens next?

Candlestick patterns are combinations of price bars that traders use to describe market behavior. Some are traditionally interpreted as signs of a possible reversal; others are associated with a continuation or a period of uncertainty. Learning to recognize them can make charts easier to read. Establishing whether they provide a reliable trading advantage requires a different kind of work. Researchers have tested candlestick patterns across different markets, timeframes and trading rules, with differing results. Some studies have found promising relationships under specific conditions; others have found little value in the patterns they examined.

In this guide, you’ll learn how to read candlesticks, explore common formations and use a printable cheat sheet. We’ll also examine what a candle can—and cannot—tell you, why patterns can look convincing in hindsight, and what the research says about turning a recognizable shape into a testable trading strategy.

First, let’s start with the building block of every pattern: the individual candlestick.

How to Read a Candlestick

A candlestick is a compact summary of how an asset’s price behaved during a chosen period. Whether you’re looking at a stock, a currency pair or another market, a standard candlestick displays four prices:

  • Open: the price at the beginning of the period.

  • High: the highest price recorded during the period.

  • Low: the lowest price recorded during the period.

  • Close: the price at the end of the period.

Together, these are known as OHLC prices: open, high, low and close.

The body and wicks

The wider part of the candle, called the body, shows the distance between the opening and closing prices. The thin lines extending above and below it are called wicks or shadows. They show how far the recorded price moved beyond the body during that period. If the price closes above its open, the candle is commonly colored green or white. In that case, the bottom of the body represents the open, and the top represents the close. If the price closes below its open, the candle is commonly colored red or black. The open is at the top of the body, and the close is at the bottom. Chart colors can be customized, so it helps to check your platform’s settings.

A simple example

Suppose a stock opens at $100, reaches a high of $106, falls as low as $98 and closes at $104. Because the stock closed above its opening price, this would typically be shown as a green candle. Its body would extend from $100 to $104, while its wicks would reach up to $106 and down to $98. The candle tells us that price traded within a range and finished higher than it began. It does not tell us exactly how the price moved between those points—or why.

Why the timeframe matters

Each candle summarizes one interval on the chart. On a five-minute chart, one candle represents five minutes of price activity. On a daily chart, it represents one day’s interval. Change the timeframe, and the chart groups the underlying price observations differently. A move that appears as one candle on an hourly chart, for example, can be broken down into several candles on a shorter timeframe.

Once you understand the individual candle, you can begin to recognize the shapes formed by one or more candles. These are what traders call candlestick patterns.

Chart reading basics

Anatomy of a candlestick

Four prices, one period. Switch between rising and falling candles to see which parts change.

One completed price interval
HighCloseOpenLow BodyWick
Rising: close above open
HighOpenCloseLow BodyWick
Falling: close below open
OpenFirst price of the interval
CloseFinal price of a completed interval
BodyRange between open and close
WicksExtensions to the high and low, when present
What each part means
OpenFirst price of the interval
CloseFinal price of a completed interval
BodyRange between open and close
WicksExtensions to the high and low, when present

What the color tells you: whether the close was above or below that candle’s own open. Colors vary across chart platforms, and a candle alone does not predict the next move.

Types of Candlestick Patterns

A single candlestick summarizes price activity over one period. A candlestick pattern is a recognizable arrangement of one or more candles that traders use to describe that activity and form an interpretation. Some patterns have memorable names, such as hammer, shooting star and morning star. They are often grouped in two ways: by the number of candles they contain and by what they are traditionally believed to suggest.

Single-, Two- and Three-Candle Patterns

Single-candle patterns are identified from the shape of one candle, although their traditional interpretation may also depend on what happened before it. Examples include the hammer, hanging man and doji.

Two-candle patterns use the relationship between two consecutive candles. For example, an engulfing pattern compares the bodies of two candles with different opening and closing directions.

Three-candle patterns consist of three candles arranged in a particular sequence. The morning star and evening star are examples commonly discussed as potential reversal formations.

Some formations use more than three candles. Across all of these groups, descriptions can differ between sources, particularly when it comes to how large a body or wick must be.

Bullish, Bearish and Neutral Patterns

Patterns are also commonly classified according to their traditional interpretation:

  • Bullish patterns are associated with the possibility of rising prices.

  • Bearish patterns are associated with the possibility of falling prices.

  • Neutral or directionally ambiguous patterns do not, on their own, suggest a clear direction. A doji, for example, is often described as a candle of indecision, though its interpretation can depend on context.

A bullish label does not mean the next candle must rise. It describes the direction a particular interpretation leads traders to look for.

Reversal and Continuation Patterns

Another distinction concerns the preceding price movement:

  • A reversal pattern is traditionally interpreted as a possible change in the direction of an existing trend.

  • A continuation pattern is traditionally interpreted as a possible pause before the existing trend resumes.

Notice the word possible. These labels describe what a formation is supposed to suggest; they do not tell us how often it has done so, whether the outcome differed from a suitable benchmark or whether a trade would have been profitable after costs.

The Same Shape Can Have Different Names

Here is where context becomes interesting. A hammer and a hanging man have similar shapes: a small body near the top of the candle and a long lower wick. Traditionally, the name and interpretation depend on where the candle appears.

After a decline, the shape may be called a hammer and interpreted as a potential bullish reversal. After an advance, a similar candle may be called a hanging man and interpreted as a potential bearish reversal.

The shape is only part of the definition. The preceding price movement is part of the traditional interpretation, too.

Explore the Patterns

The interactive atlas below brings these formations together in one place. Use it to compare their shapes, see how they are traditionally classified and review the context commonly associated with each one. You can also use the printable cheat sheet as a quick reference.

As you explore, keep one distinction in mind: identifying a pattern and establishing that it has predictive value are two separate steps. Later in this guide, we’ll look at how researchers have tested that second claim.

FINANSIFIED · AUDITED ILLUSTRATIONS

Candlestick Pattern Atlas

37 selected patterns and related candle structures. Compare traditional labels with the actual candle geometry and the market context they require.

How to read: Green = close above open · Red = close below open · Gray = preceding trend context (not counted). Descriptions are not trading signals.

Filter patterns and choose print format
Traditional interpretation
Pattern length
Print selection

Print only the patterns: Choose Quick Sheet (16) or Full Reference (37), then use Ctrl+P / Cmd+P. Print CSS excludes the surrounding article. The optional interactive edition also offers direct Print buttons.

37 audited illustrationsScroll this list · Tap a row to enlarge its candles

Candlestick patterns · Finansified reference

Schematic OHLC illustrations, not validated forecasts. Green candles close above open; red candles close below open; gray candles show prior context and are not counted in the pattern.

HammerBullish · 1 candle · ReversalShape: Small body near the high, long lower wick and little upper wick.
Context: After a decline.
Note: The body may be green or red. The same shape is called Hammer after a decline or Hanging Man after an advance; the surrounding trend is essential.

Price context: After a decline.

Recognizing the shape: Small body near the high, long lower wick and little upper wick.

Definition boundary: The body may be green or red. The same shape is called Hammer after a decline or Hanging Man after an advance; the surrounding trend is essential.

Gray candles are preceding context and not part of the counted pattern. Reference: StockCharts / TA-Lib.

Hanging manBearish · 1 candle · ReversalShape: Small body near the high, long lower wick and little upper wick.
Context: After an advance.
Note: The body may be green or red. The same shape is called Hammer after a decline or Hanging Man after an advance; the surrounding trend is essential.

Price context: After an advance.

Recognizing the shape: Small body near the high, long lower wick and little upper wick.

Definition boundary: The body may be green or red. The same shape is called Hammer after a decline or Hanging Man after an advance; the surrounding trend is essential.

Gray candles are preceding context and not part of the counted pattern. Reference: StockCharts / TA-Lib.

Inverted hammerBullish · 1 candle · ReversalShape: Small body near the low with a long upper wick.
Context: After a decline.
Note: Definitions vary; illustrations are schematic.

Price context: After a decline.

Recognizing the shape: Small body near the low with a long upper wick.

Definition boundary: Definitions vary; illustrations are schematic.

Gray candles are preceding context and not part of the counted pattern. Reference: StockCharts / TA-Lib.

Shooting starBearish · 1 candle · ReversalShape: Small body near the low with a long upper wick.
Context: After an advance.
Note: Similar in shape to an inverted hammer; context affects the traditional name.

Price context: After an advance.

Recognizing the shape: Small body near the low with a long upper wick.

Definition boundary: Similar in shape to an inverted hammer; context affects the traditional name.

Gray candles are preceding context and not part of the counted pattern. Reference: StockCharts / TA-Lib.

DojiContextual · 1 candle · IndecisionShape: Open and close are equal or very close, with wicks of varying lengths.
Context: May occur in many market conditions.
Note: Definitions vary; illustrations are schematic.

Price context: May occur in many market conditions.

Recognizing the shape: Open and close are equal or very close, with wicks of varying lengths.

Definition boundary: Definitions vary; illustrations are schematic.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Long-legged dojiContextual · 1 candle · IndecisionShape: Open and close are nearly equal with long upper and lower wicks.
Context: May appear during a wide-ranging period.
Note: Definitions vary; illustrations are schematic.

Price context: May appear during a wide-ranging period.

Recognizing the shape: Open and close are nearly equal with long upper and lower wicks.

Definition boundary: Definitions vary; illustrations are schematic.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Dragonfly dojiContextual · 1 candle · IndecisionShape: Open and close are near the high with a long lower wick.
Context: Sometimes interpreted as a possible reversal, depending on trend.
Note: A descriptive doji shape. The direction is not determined by this candle alone; location and price context matter.

Price context: Sometimes interpreted as a possible reversal, depending on trend.

Recognizing the shape: Open and close are near the high with a long lower wick.

Definition boundary: A descriptive doji shape. The direction is not determined by this candle alone; location and price context matter.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Gravestone dojiContextual · 1 candle · IndecisionShape: Open and close are near the low with a long upper wick.
Context: Sometimes interpreted as a possible reversal, depending on trend.
Note: A descriptive doji shape. The direction is not determined by this candle alone; location and price context matter.

Price context: Sometimes interpreted as a possible reversal, depending on trend.

Recognizing the shape: Open and close are near the low with a long upper wick.

Definition boundary: A descriptive doji shape. The direction is not determined by this candle alone; location and price context matter.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Spinning topContextual · 1 candle · IndecisionShape: A small body with noticeable upper and lower wicks.
Context: Can appear in an advance, decline or range.
Note: Definitions vary; illustrations are schematic.

Price context: Can appear in an advance, decline or range.

Recognizing the shape: A small body with noticeable upper and lower wicks.

Definition boundary: Definitions vary; illustrations are schematic.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Bullish marubozuBullish · 1 candle · Strong directional candleShape: A long rising body with little or no upper or lower wick.
Context: Often read as strong upward price movement; context matters.
Note: A long candle body describes movement during the period. It is not inherently a continuation forecast.

Price context: Often read as strong upward price movement; context matters.

Recognizing the shape: A long rising body with little or no upper or lower wick.

Definition boundary: A long candle body describes movement during the period. It is not inherently a continuation forecast.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Bearish marubozuBearish · 1 candle · Strong directional candleShape: A long falling body with little or no upper or lower wick.
Context: Often read as strong downward price movement; context matters.
Note: A long candle body describes movement during the period. It is not inherently a continuation forecast.

Price context: Often read as strong downward price movement; context matters.

Recognizing the shape: A long falling body with little or no upper or lower wick.

Definition boundary: A long candle body describes movement during the period. It is not inherently a continuation forecast.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Bullish engulfingBullish · 2 candles · ReversalShape: A rising real body covers the prior falling real body.
Context: Traditionally after a decline.
Note: Definitions vary; illustrations are schematic.

Price context: Traditionally after a decline.

Recognizing the shape: A rising real body covers the prior falling real body.

Definition boundary: Definitions vary; illustrations are schematic.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Bearish engulfingBearish · 2 candles · ReversalShape: A falling real body covers the prior rising real body.
Context: Traditionally after an advance.
Note: Definitions vary; illustrations are schematic.

Price context: Traditionally after an advance.

Recognizing the shape: A falling real body covers the prior rising real body.

Definition boundary: Definitions vary; illustrations are schematic.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Bullish haramiBullish · 2 candles · ReversalShape: A small second real body sits inside a larger preceding falling body.
Context: Traditionally after a decline.
Note: Definitions vary; illustrations are schematic.

Price context: Traditionally after a decline.

Recognizing the shape: A small second real body sits inside a larger preceding falling body.

Definition boundary: Definitions vary; illustrations are schematic.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Bearish haramiBearish · 2 candles · ReversalShape: A small second real body sits inside a larger preceding rising body.
Context: Traditionally after an advance.
Note: Definitions vary; illustrations are schematic.

Price context: Traditionally after an advance.

Recognizing the shape: A small second real body sits inside a larger preceding rising body.

Definition boundary: Definitions vary; illustrations are schematic.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Piercing lineBullish · 2 candles · ReversalShape: A falling candle is followed by a rising candle that opens below its low and closes above the midpoint of its body, but below its open.
Context: Traditionally after a decline.
Note: Gap requirements vary; strict definitions can occur less often in continuous markets.

Price context: Traditionally after a decline.

Recognizing the shape: A falling candle is followed by a rising candle that opens below its low and closes above the midpoint of its body, but below its open.

Definition boundary: Gap requirements vary; strict definitions can occur less often in continuous markets.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Dark cloud coverBearish · 2 candles · ReversalShape: A rising candle is followed by a falling candle that opens above its high and closes below the midpoint of its body, but above its open.
Context: Traditionally after an advance.
Note: Gap requirements vary; strict definitions can occur less often in continuous markets.

Price context: Traditionally after an advance.

Recognizing the shape: A rising candle is followed by a falling candle that opens above its high and closes below the midpoint of its body, but above its open.

Definition boundary: Gap requirements vary; strict definitions can occur less often in continuous markets.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Tweezer bottomBullish · 2 candles · ReversalShape: Two nearby candles have matching or nearly matching lows.
Context: Traditionally after a decline.
Note: How close the lows must be is not universally standardized.

Price context: Traditionally after a decline.

Recognizing the shape: Two nearby candles have matching or nearly matching lows.

Definition boundary: How close the lows must be is not universally standardized.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Tweezer topBearish · 2 candles · ReversalShape: Two nearby candles have matching or nearly matching highs.
Context: Traditionally after an advance.
Note: How close the highs must be is not universally standardized.

Price context: Traditionally after an advance.

Recognizing the shape: Two nearby candles have matching or nearly matching highs.

Definition boundary: How close the highs must be is not universally standardized.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Morning starBullish · 3 candles · ReversalShape: A long falling candle, small-bodied middle candle and rising candle closing well into the first body.
Context: Traditionally after a decline.
Note: Some definitions require gaps around the middle body; others use looser variants.

Price context: Traditionally after a decline.

Recognizing the shape: A long falling candle, small-bodied middle candle and rising candle closing well into the first body.

Definition boundary: Some definitions require gaps around the middle body; others use looser variants.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Evening starBearish · 3 candles · ReversalShape: A long rising candle, small-bodied middle candle and falling candle closing well into the first body.
Context: Traditionally after an advance.
Note: Some definitions require gaps around the middle body; others use looser variants.

Price context: Traditionally after an advance.

Recognizing the shape: A long rising candle, small-bodied middle candle and falling candle closing well into the first body.

Definition boundary: Some definitions require gaps around the middle body; others use looser variants.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Morning doji starBullish · 3 candles · ReversalShape: A morning star variation whose middle candle is a doji.
Context: Traditionally after a decline.
Note: Gap and body-size rules vary across sources.

Price context: Traditionally after a decline.

Recognizing the shape: A morning star variation whose middle candle is a doji.

Definition boundary: Gap and body-size rules vary across sources.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Evening doji starBearish · 3 candles · ReversalShape: An evening star variation whose middle candle is a doji.
Context: Traditionally after an advance.
Note: Gap and body-size rules vary across sources.

Price context: Traditionally after an advance.

Recognizing the shape: An evening star variation whose middle candle is a doji.

Definition boundary: Gap and body-size rules vary across sources.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Three white soldiersBullish · 3 candles · ReversalShape: Three consecutive long rising candles with progressively higher closes.
Context: Traditionally after a decline.
Note: Traditionally a reversal after a preceding move; in other contexts the same sequence may be described as continuation.

Price context: Traditionally after a decline.

Recognizing the shape: Three consecutive long rising candles with progressively higher closes.

Definition boundary: Traditionally a reversal after a preceding move; in other contexts the same sequence may be described as continuation.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Three black crowsBearish · 3 candles · ReversalShape: Three consecutive long falling candles with progressively lower closes.
Context: Traditionally after an advance.
Note: Traditionally a reversal after a preceding move; in other contexts the same sequence may be described as continuation.

Price context: Traditionally after an advance.

Recognizing the shape: Three consecutive long falling candles with progressively lower closes.

Definition boundary: Traditionally a reversal after a preceding move; in other contexts the same sequence may be described as continuation.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Three inside upBullish · 3 candles · ReversalShape: A bullish harami followed by a rising candle closing above the first candle’s open.
Context: Traditionally after a decline.
Note: Some sources use a different confirmation threshold.

Price context: Traditionally after a decline.

Recognizing the shape: A bullish harami followed by a rising candle closing above the first candle’s open.

Definition boundary: Some sources use a different confirmation threshold.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Three inside downBearish · 3 candles · ReversalShape: A bearish harami followed by a falling candle closing below the first candle’s open.
Context: Traditionally after an advance.
Note: Some sources use a different confirmation threshold.

Price context: Traditionally after an advance.

Recognizing the shape: A bearish harami followed by a falling candle closing below the first candle’s open.

Definition boundary: Some sources use a different confirmation threshold.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Bullish harami crossBullish · 2 candles · ReversalShape: A long falling body followed by a doji whose body lies within it.
Context: Traditionally after a decline.
Note: Definitions vary; illustrations are schematic.

Price context: Traditionally after a decline.

Recognizing the shape: A long falling body followed by a doji whose body lies within it.

Definition boundary: Definitions vary; illustrations are schematic.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Bearish harami crossBearish · 2 candles · ReversalShape: A long rising body followed by a doji whose body lies within it.
Context: Traditionally after an advance.
Note: Definitions vary; illustrations are schematic.

Price context: Traditionally after an advance.

Recognizing the shape: A long rising body followed by a doji whose body lies within it.

Definition boundary: Definitions vary; illustrations are schematic.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Rising three methodsBullish · 5 candles · ContinuationShape: A long rising candle, a short countertrend sequence inside its range, then a long rising candle with a higher close.
Context: Traditionally within an advance.
Note: The middle section often has three candles; descriptions differ on exact overlap.

Price context: Traditionally within an advance.

Recognizing the shape: A long rising candle, a short countertrend sequence inside its range, then a long rising candle with a higher close.

Definition boundary: The middle section often has three candles; descriptions differ on exact overlap.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Falling three methodsBearish · 5 candles · ContinuationShape: A long falling candle, a short countertrend sequence inside its range, then a long falling candle with a lower close.
Context: Traditionally within a decline.
Note: The middle section often has three candles; descriptions differ on exact overlap.

Price context: Traditionally within a decline.

Recognizing the shape: A long falling candle, a short countertrend sequence inside its range, then a long falling candle with a lower close.

Definition boundary: The middle section often has three candles; descriptions differ on exact overlap.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: StockCharts / TA-Lib.

Doji star (gap down)Contextual · 2 candles · Two-candle starShape: A doji whose real body gaps below the first long falling candle’s real body.
Context: A long falling candle followed by a lower doji; preceding decline matters.
Note: Shown as a down-gap arrangement. Doji Star can also occur as a gap-up arrangement following a rising candle. Its traditional bias depends on context; its predictive value is not guaranteed.

Price context: A long falling candle followed by a lower doji; preceding decline matters.

Recognizing the shape: A doji whose real body gaps below the first long falling candle’s real body.

Definition boundary: Shown as a down-gap arrangement. Doji Star can also occur as a gap-up arrangement following a rising candle. Its traditional bias depends on context; its predictive value is not guaranteed.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: TA-Lib CDLDOJISTAR.

Inside bar (range containment)Contextual · 2 candles · Descriptive price structureShape: The second candle’s entire high-to-low range lies within the first candle’s high-to-low range.
Context: Can appear in a trend or a trading range.
Note: This is a general price-action structure, not a separate named TA-Lib candlestick recognition function. It carries no necessary directional forecast.

Price context: Can appear in a trend or a trading range.

Recognizing the shape: The second candle’s entire high-to-low range lies within the first candle’s high-to-low range.

Definition boundary: This is a general price-action structure, not a separate named TA-Lib candlestick recognition function. It carries no necessary directional forecast.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: OHLC range definition.

Tri-star (gap-up example)Contextual · 3 candles · Three-doji reversal familyShape: Three doji, with the middle doji gapped above its neighbors in the depicted form.
Context: Illustrates the bearish, gap-up configuration; a mirrored bullish form exists.
Note: TA-Lib treats gap-up and gap-down Tri-Star variants as bearish and bullish respectively; the preceding trend still requires assessment.

Price context: Illustrates the bearish, gap-up configuration; a mirrored bullish form exists.

Recognizing the shape: Three doji, with the middle doji gapped above its neighbors in the depicted form.

Definition boundary: TA-Lib treats gap-up and gap-down Tri-Star variants as bearish and bullish respectively; the preceding trend still requires assessment.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: TA-Lib CDLTRISTAR.

Mat holdBullish · 5 candles · ContinuationShape: Long rising candle, small gapped-up falling candle, two short pullback candles, then a rising candle that clears the intervening highs.
Context: Traditionally after an advance.
Note: This is one idealized five-candle bullish variant; automated definitions may use historical body-size and penetration thresholds.

Price context: Traditionally after an advance.

Recognizing the shape: Long rising candle, small gapped-up falling candle, two short pullback candles, then a rising candle that clears the intervening highs.

Definition boundary: This is one idealized five-candle bullish variant; automated definitions may use historical body-size and penetration thresholds.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: TA-Lib CDLMATHOLD.

Bullish breakawayBullish · 5 candles · ReversalShape: An extended five-candle decline with a real-body gap early in the sequence, followed by a rising fifth candle closing back inside that gap.
Context: Traditionally after a decline.
Note: Schematic gap-dependent version; rare in many markets and timeframes. Traditional direction is not an empirical guarantee.

Price context: Traditionally after a decline.

Recognizing the shape: An extended five-candle decline with a real-body gap early in the sequence, followed by a rising fifth candle closing back inside that gap.

Definition boundary: Schematic gap-dependent version; rare in many markets and timeframes. Traditional direction is not an empirical guarantee.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: TA-Lib CDLBREAKAWAY.

Bearish breakawayBearish · 5 candles · ReversalShape: An extended five-candle advance with a real-body gap early in the sequence, followed by a falling fifth candle closing into that gap.
Context: Traditionally after an advance.
Note: Schematic gap-dependent version; rare in many markets and timeframes. Traditional direction is not an empirical guarantee.

Price context: Traditionally after an advance.

Recognizing the shape: An extended five-candle advance with a real-body gap early in the sequence, followed by a falling fifth candle closing into that gap.

Definition boundary: Schematic gap-dependent version; rare in many markets and timeframes. Traditional direction is not an empirical guarantee.

The drawing is schematic; relative sizes and gap rules may differ by definition. Reference: TA-Lib CDLBREAKAWAY.

Interpretation is not prediction. A valid OHLC schematic does not establish that an algorithm would recognize a pattern in real data. Definitions may require prior volatility, body-size thresholds, specific gaps and trend context. Five-candle gaps are less common in continuous FX trading than in some exchange-traded markets. No claim of reliable predictive power is implied.

What Are Candlestick Patterns Supposed to Tell Us?

A candlestick pattern gives traders a way to describe a sequence of price movements. Its name, however, usually comes with a story about what those movements might mean.

Consider the hammer. After a decline, a candle with a small body near the top and a long lower wick is often interpreted as a sign that the downward move may be losing strength. In the traditional account, sellers pushed the price lower, but buying interest helped bring it back up before the candle closed.

That is a plausible interpretation. What we can observe directly is more limited: the recorded price reached a low and ended the period above it. The candle does not identify the traders involved, reveal their motives or establish that a reversal will follow.

Why the Preceding Trend Matters

For many patterns, the conventional interpretation depends on what happened before the candle formed.

As we saw in the atlas, a hammer and a hanging man can look very similar. After a decline, the shape is traditionally called a hammer and viewed as a potential bullish reversal. After an advance, a similar shape is called a hanging man and viewed as a potential bearish reversal.

The distinction makes sense within the language of pattern analysis: a reversal has to reverse something. But it also introduces a practical question: how much of a decline or advance counts as a trend?

One trader might look at the previous few candles; another might use several weeks of prices or a mathematical rule. Different choices can lead them to identify different patterns on the same chart.

What Does “Confirmation” Mean?

Traders often look for confirmation after a pattern appears: an additional price movement or condition they believe supports its interpretation.

For example, after a hammer, a trader might wait for a later candle to close above the hammer’s high before considering a bullish trade. Another might require the price to rise by a certain amount or use an indicator as an additional filter.

Waiting for confirmation changes the decision. The trader has more information than they had when the hammer closed, but may also enter at a different price or miss some moves. It also changes what needs to be tested: the question becomes whether the pattern combined with that particular confirmation rule has been useful—not whether the pattern alone has worked.

Confirmation does not guarantee a better outcome. A rule intended to filter out unsuccessful signals can itself produce misleading results, so it needs to be evaluated on data beyond the examples used to select it.

A Description Is Not Yet an Explanation

The language surrounding candlesticks can make a chart feel like a window into the minds of market participants: buyers took control, sellers gave up, or the market became indecisive.

Those stories can help explain why traders pay attention to a formation. But the same recorded price movement may arise from different combinations of orders, liquidity conditions and information. A candle summarizes the prices; it does not tell us which explanation is correct.

That distinction leads to a deeper question: what evidence would connect a pattern to a subsequent move—and what would it take to show that the relationship is causal?

Is There a Cause-and-Effect Relationship Behind Candlestick Patterns?

Imagine watching a price fall sharply, recover and close near the top of its range. On a candlestick chart, that movement might leave a long lower wick and a small body.

What caused the recovery?

Perhaps buyers entered at lower prices. Perhaps selling slowed, liquidity changed, or new information prompted traders to adjust their positions. The candle can show us the resulting prices, but it cannot tell us which explanation is correct.

To see why that matters, it helps to separate four ideas that are often blended together.

Observation: What Happened?

The first statement describes what the chart records:

The price reached a low during the period and closed substantially above it.

That is an observation based on the candle’s price data. If the shape meets a chosen definition and appears after a decline, a trader might identify it as a hammer.

Interpretation: What Might It Mean?

The next step adds an explanation:

The recovery may indicate that buying interest increased, selling pressure eased, or both.

That is a possible interpretation of the price movement. It is not something the candle alone can verify. To investigate what happened to orders or liquidity during the interval, we would need information beyond the four prices summarized by the candle.

Prediction: What Might Happen Next?

A trader could turn that interpretation into a forecast:

After a hammer appears in a declining market, subsequent returns will tend to be higher than an appropriate comparison.

Now we have a claim that can be tested. But the details matter: How is the decline defined? How long after the pattern do we measure returns? Which market are we studying, and what are we comparing the result with?

A handful of examples where prices rose afterward cannot answer those questions. We would also need to examine the cases where the pattern appeared and prices fell or moved very little.

Causal Explanation: Why Would It Happen?

Suppose a study finds that prices tended to rise after a precisely defined hammer in a particular market. That would be evidence of an association under the conditions studied.

It would not, by itself, establish why the relationship occurred.

One possibility is that the pattern reflects a change in trading activity. Another is that both the pattern and the later return are related to some other condition, such as volatility or information released to the market. The apparent relationship might also be sensitive to the sample or the way the pattern was selected.

A causal explanation requires evidence that helps distinguish between these alternatives.

Could There Be a Market Mechanism?

Yes. Financial prices are shaped by real processes: information arrives, participants submit orders, liquidity changes and dealers manage their positions. There is no reason to assume that every relationship involving past prices must be meaningless.

Research into foreign exchange market structure, for example, has examined the relationship between order flow—a measure of buying and selling pressure derived from transactions—and exchange-rate movements. That work helps explain why trading activity can matter for prices. It does not establish that a particular candlestick pattern reliably predicts the next move.

It is also possible to imagine a feedback effect: if enough traders watch the same formation and respond to it, their actions could affect prices. Whether that occurs for a specific pattern, and whether any resulting effect survives trading costs, is an empirical question.

The Question to Ask

A candlestick formation can be a useful starting point for a hypothesis. The important step is to be clear about which claim is being made:

  • Observation: What did the recorded price do?

  • Interpretation: What might explain that movement?

  • Prediction: What do we expect to happen afterward?

  • Causal explanation: What evidence supports the proposed reason?

These questions are connected, but answering one does not automatically answer the others.

And there is another complication: even the price movement inside a candle is less complete than it appears. Two very different paths can produce exactly the same open, high, low and close.

What a candlestick leaves out

Same candle, different story

Two price paths can reach the same high and low in a different order, yet produce exactly the same candle.

Highlight paths

Two possible price paths

Price ↑ · time → · illustrative observations

11010090OpenClose 11010090OpenClose
Path A: Open → High → Low → Close
Path B: Open → Low → High → Close

One identical candle

The same four summary prices

HighCloseOpenLow
Open100
High110
Low90
Close104

The key limitation: OHLC captures where a period opened, closed and reached its extremes, but not whether the high came before the low or what happened between those points.

Hypothetical paths, not observed trades. Shorter-timeframe data may reveal more of the sequence, but a single OHLC candle cannot reconstruct it.

Why the Same Market Can Produce Different Candlesticks

Two traders open daily charts of EUR/USD on different platforms. One sees a small-bodied candle with a long lower wick. The other sees a larger body and a shorter wick.

They may both be looking at the same currency pair on the same date. So why do the candles look different?

A candle is constructed from the prices included in a particular interval. Change the interval—or the prices used to build it—and you can change the candle.

The Timeframe Changes the Picture

A daily candle combines the price activity assigned to one day’s interval. A four-hour chart divides that activity into shorter periods, producing several candles instead.

A long wick on the daily chart might contain a sequence of smaller declines, recoveries and pauses on the four-hour chart. Patterns can therefore appear on one timeframe but not another.

That does not tell us which chart is “correct.” They summarize price activity at different levels of detail. Whether a pattern identified on either timeframe has predictive value is a separate question.

The Daily Closing Time Changes the Candle

A daily candle needs a starting and ending point. In forex, where trading continues across time zones during the trading week, platforms may use different boundaries to define a trading day.

For example, one chart might construct daily candles from midnight to midnight UTC. Another might use 5 p.m. New York time as the daily closing point.

Those candles cover different 24-hour windows. A price move included near the end of one daily candle could appear near the beginning of another. As a result, their opens, highs, lows and closes can differ—even when they are built from the same underlying price observations.

There is an additional detail: New York observes daylight saving time. A 5 p.m. New York close corresponds to 22:00 UTC during standard time and 21:00 UTC during daylight saving time.

Changing the time labels displayed on a chart does not necessarily change how its candles are constructed. The important setting is the one that controls the candle’s actual start and end.

The Price Source Can Make a Difference

Forex charts may also be built from different prices:

  • Bid: the quoted price at which a trader can generally sell.

  • Ask: the quoted price at which a trader can generally buy.

  • Midpoint: the price halfway between the bid and ask.

Because the bid and ask differ by the spread, candles constructed from them can have different recorded values. A midpoint chart can look different again. Small differences may matter when a pattern’s definition depends on two prices being equal, nearly equal or just beyond a particular level.

Prices can also differ between data providers. The forex market is decentralized, so feeds do not necessarily contain identical quotes, timestamps or aggregation choices.

Does That Make the Patterns Meaningless?

Not necessarily. A pattern could still contain useful information under a particular set of chart settings. But if a trader—or a researcher—claims that a formation has predictive value, the settings are part of the claim.

A reproducible test should therefore specify the data source, timeframe, candle boundaries and price series used to identify the pattern. Otherwise, someone trying to repeat the analysis may not be looking at the same candles.

The following interactive makes the effect of one of those choices visible: what happens when we keep the underlying price observations and change where the daily candle begins and ends?

Chart construction · Interactive comparison

Change the clock. Change the candle.

The same hourly prices can produce different daily candles when the cutoff time changes. Compare midnight UTC with the 5 p.m. New York convention.

One underlying price series

193 simulated hourly observations

6 Jul 2026 · 00:00 UTC14 Jul · 00:00 UTC
Show boundaries

Two ways to build daily candles

Same price scale · Six 24-hour windows each

Midnight UTC

00:00 UTC cutoff

1 body-only engulfing match

New York close

17:00 EDT = 21:00 UTC

0 body-only engulfing matches

What changed? Only the daily cutoffs. A two-candle shape appears in one aggregation and disappears in the other. The shape alone is not a validated trading signal.

Method, assumptions & limitations

Illustrative data: This is a synthetic, uninterrupted hourly series, not historical EUR/USD prices. It deliberately includes weekend hours; a real spot-FX feed has trading breaks and may have missing observations.

Aggregation: Each candle has 24 one-hour intervals. Its open is the first observation, close the observation 24 hours later, and sampled high/low the extrema of all 25 hourly endpoint observations. Adjacent candles share a boundary observation. Intrahour highs and lows are unknown. Real platforms can use different quote/session rules.

Timing: The sample starts 6 July 2026. In July, 5 p.m. New York is 21:00 UTC; in standard time it is 22:00 UTC. Both panels contain six complete windows, but the New York sample starts 21 hours later, so the displayed overall intervals are offset.

Pattern marker: The ring denotes only a falling real body followed by a rising real body covering the first body (inclusive boundaries). It does not check the preceding trend, strict gap rules, confirmation, or predictive value.

Why Candlestick Patterns Can Look So Convincing

Imagine scrolling through a historical chart and finding a hammer just before a sharp rally. A little farther back, you spot another one followed by a similar move.

The examples are real. The formations are recognizable. And the reversals appear to support the pattern’s traditional interpretation.

But there is a question the chart cannot answer by showing us those examples alone: how many hammers appeared without a meaningful rally afterward?

Recognizing a Shape Is Not the Same as Recognizing a Signal

Human beings are good at finding patterns. That ability helps us make sense of complicated information—from recognizing a familiar face to noticing a change in a market chart.

But recognizable shapes can also appear in sequences generated by chance. Imagine tossing a fair coin repeatedly: a run of five heads may look remarkable, yet it can occur without the coin becoming more likely to land heads on the next toss.

A chart generated from random price changes can contain runs, turning points and candle shapes that resemble familiar formations. The appearance of a pattern does not, by itself, tell us whether it contains information about what comes next.

That does not mean financial markets are always random, or that every apparent relationship is an illusion. It means that visual recognition is a starting point for investigation, not the end of it.

The Advantage of Looking Backward

Historical charts have an important feature that live charts do not: we already know what happened next.

Once a reversal is visible, it becomes easier to find a candle near the turning point and tell a convincing story about it. Meanwhile, similar candles elsewhere on the chart may have been followed by declines, sideways movement or only a brief rise.

A fair evaluation needs both groups. If we count only the examples that worked, we cannot tell how often the pattern appeared without the expected result.

For example, suppose we find 100 instances of a defined pattern and the price rises afterward in 55 cases. That sounds encouraging—until we ask what happened following comparable periods without the pattern. If prices also rose in 55 out of 100 comparable cases, the pattern may not have added useful directional information in that example.

The numbers here are illustrative, not a measured success rate. The comparison matters because a pattern can appear to work simply because the market was already more likely to rise over the chosen period.

When Searching Creates Discoveries

There is another reason attractive results can appear: researchers and traders have many choices to make.

They can examine dozens of patterns, markets, timeframes, holding periods and filters. When enough combinations are tested, some may look unusually successful by chance. If only the strongest result is reported, it can appear more persuasive than it would if readers could see the full set of tests.

This problem is known as data snooping. It does not mean a discovered result is automatically false. It means the process used to find it matters when deciding how much confidence to place in it.

One way to investigate further is to define the pattern and test rules clearly, then evaluate them on data that were not used to select or refine those rules. We’ll look at that process—and its limitations—later in the guide.

Put Your Pattern-Spotting to the Test

The interactive exercise below lets you make predictions before seeing what comes next. It is an informal illustration, not a scientific test of your trading ability.

As you use it, notice how compelling a chart can feel when a familiar shape appears—and how different the question becomes when you must make the prediction before the next movement is revealed.

Interactive experiment · Pattern perception

Can you see what comes next?

Study the price path, predict the next move, then reveal what happened. Try ten rounds and see how easily patterns can deceive.

CHOOSE A DATASET

ROUND 01 / 100 correct
SIMULATED RANDOM WALKLatest: 100.00
12 observationsNext value hidden

Will the next observation be higher or lower?

View the observations as numbers

Methodology and limitations

Simulation: Each round is generated in advance using independent, symmetric positive or negative changes. The next sign has a 50% chance of being positive, even after a convincing-looking chart. Results can vary substantially over ten rounds.

Historical mode: Uses 51 Federal Reserve H.10 EUR/USD reference observations from 2 January to 14 March 2024 (US dollars per euro). Ten eligible windows are sampled without replacement, although their price histories can overlap. Equal successive values are excluded. Calendar gaps and holidays remain; date labels appear only after a prediction.

Important limitations: The historical mode is a short, selected period. These are daily reference observations, not executable quotes or candlesticks. A historical majority-direction score is calculated in hindsight and is not a strategy anyone could have followed beforehand. The exercise does not test trading profitability, the efficiency of markets, or anyone’s real forecasting skill.

Original data: Federal Reserve · H.10 EUR/USD history. The experiment runs locally in your browser; it sends no answers to a server.

This interactive experiment requires JavaScript to generate the price paths, reveal each next observation and calculate the results. The methodology and Federal Reserve source are available above.
FINANSIFIED

Do Candlestick Patterns Actually Work? What the Research Finds

If candlestick patterns were reliable forecasting tools, we would expect researchers to find that clearly defined formations are followed by distinguishable price behavior—and that any apparent advantage survives a careful test. Researchers have examined that question in stocks, currencies and cryptocurrencies. Their findings do not fit neatly into a single “yes” or “no.”

Some studies have reported promising results for particular patterns or strategies. Others have found little value in the formations they tested. Even studies examining similar markets have reached different conclusions when they used different rules. Candlestick patterns are one part of a broader question about whether technical analysis works. The important question is not simply “Do candlestick patterns work?” It is:

Which pattern, defined how, in which market, over what period—and useful for what purpose?

U.S. Stocks: A Challenge to Candlestick Strategies

In a 2006 study, Marshall, Young and Rose examined candlestick trading strategies using Dow Jones Industrial Average component stocks from 1992 to 2002.

The researchers used a statistical technique called bootstrapping to help evaluate the results against simulated price series. Their conclusion was that the tested candlestick strategies did not create value for investors in that sample.

That is a meaningful challenge to broad claims about candlestick trading. It is not a test of every possible pattern, market or strategy. The result applies to the formations, stocks, period and trading rules the researchers examined.

A later U.S. stock study illustrates why those details matter.

The Same Market, a Different Exit Rule

In 2015, Lu, Chen and Hsu investigated candlestick strategies using Dow stocks and compared different ways of defining trends and exiting trades.

They reported profitable results for eight tested three-candle reversal patterns under one holding strategy, after an assumed transaction cost and an adjustment intended to address data snooping. Under another holding strategy, the patterns were not profitable.

The difference was not simply the name of the pattern. It was also what the hypothetical investor did after entering the trade.

That is an important distinction. A candlestick formation may describe a possible entry condition, but an investment result also depends on the exit, position, costs and benchmark. A study of one complete strategy cannot automatically validate—or invalidate—every strategy that uses the same formation.

Taiwan: Some Positive Results

In a study of Taiwanese stocks from 1992 to 2009, Tsung-Hsun Lu examined a set of single-candle patterns.

The study reported four profitable signals after its assumed transaction costs. The author also used bootstrap analysis, subsamples and an out-of-sample exercise to investigate the robustness of the findings.

Those checks make the positive results worth examining. They do not establish that the same formations will be profitable in other markets or future periods. The pattern definitions, stock market, dates and trading assumptions remain part of the finding.

Thailand: Most Tested Reversal Patterns Fell Short

A 2017 study by Tharavanij, Siraprapasiri and Rajchamaha examined reversal patterns in Thai SET50 stocks from 2006 to 2016.

The researchers tested several holding periods and two exit approaches. They also examined whether adding filters based on indicators such as RSI improved the results.

They found that the average returns following most of the tested patterns were not statistically distinguishable from zero. Some individual results were statistically significant, but the authors noted that these returns were accompanied by substantial variability. The tested filters generally did not improve profitability or directional accuracy.

This does not prove that adding a filter can never help. It does show that an extra indicator should not be assumed to improve a pattern simply because it makes the trading rule look more selective.

Forex: Engulfing Patterns and Trading Costs

In a 2020 paper, Ahmed Alanazi studied bullish and bearish engulfing patterns across 24 currency pairs using data from 2000 to 2018.

The author reported evidence of profitability for the tested approach, particularly among major currency pairs, while also emphasizing the importance of transaction costs.

The paper describes an in-sample test using daily candles and an out-of-sample exercise using four-hour candles. That is a useful additional check, although the change in timeframe means readers should pay close attention to exactly what is being compared. It should not be treated as proof that every engulfing pattern is profitable across forex markets.

Cryptocurrency: Predictability Is Not the Same as Profitability

In a 2026 study, Stefanie Moser and Alexander Brauneis examined 55 reversal patterns in hourly cryptocurrency data from July 2018 to January 2022.

The dataset covered nearly 400 cryptocurrencies, approximately 2,000 trading pairs and 36 exchanges. The authors reported that some patterns were followed by statistically significant price movements in the directions traditionally associated with them. They also used a statistical procedure designed to address the problem of testing many patterns.

This is positive evidence that some formations contained information about subsequent returns in the cryptocurrency data studied.

But the researchers’ core comparison did not include trading costs. The paper also discusses the practical difficulty of trading at the closing price of the candle that completes a pattern. A small predicted movement may be valuable in a statistical test and still be insufficient to cover fees, spreads or other execution costs.

That is not a reason to dismiss the finding. It is a reason to describe precisely what the finding establishes.

What Can We Conclude?

The studies above are a selection of the research, not a complete worldwide count. They show why a simple verdict can be misleading:

  • Some tested patterns or strategies have produced positive results under particular conditions.

  • Other tests have found little value, or results that weaken after costs and statistical adjustments.

  • The way a study defines a pattern, selects a market and constructs its trading rules can change the conclusion.

  • Evidence of a relationship with future prices is not automatically evidence of a profitable, executable strategy.

A fair assessment must leave room for both positive and negative evidence. It must also ask whether a result is large, robust and practical enough to matter.

That brings us to the next question: how can two serious studies examine candlesticks and arrive at different answers?

Evidence desk · Candlestick research

What have researchers found?

Explore seven published studies across equities, forex, and cryptocurrencies. Compare the reported findings, then inspect the assumptions behind them.

Selected studies, not a worldwide census. Different tests, markets and costs cannot be combined into a success rate.

  1. 2006 · Stocks · United States

    Candlestick technical trading strategies: Can they create value for investors?

    Marshall, Young & Rose · Journal of Banking & Finance

    Reported result: limited evidence

    The authors found no value for the tested strategies in their DJIA stock sample.

    Original publication ↗
    Methods & limitations
    Sample
    Dow Jones Industrial Average stocks, 1992–2002
    Frequency
    Daily
    Patterns examined
    Candlestick trading strategies; see paper for the exact rules.
    Transaction costs
    Details require further verification
    Methods and benchmarks
    Bootstrap comparison with simulated OHLC series.
    Out-of-sample assessment
    Not established in this summary
    Multiple-testing treatment
    Bootstrap-based tests; how the analysis accounts for multiple searches requires further review.
    Verification status
    Publisher text and abstract reviewed; complete methodology not independently verified.
    Limitations
    The conclusion is confined to the instruments, historical period, definitions and trading rules tested.
  2. 2013 · Stocks · United States

    The intra-day performance of market timing strategies and trading systems based on Japanese candlesticks

    Duvinage, Mazza & Petitjean · Quantitative Finance

    Reported result: mixed findings

    Some rules showed intraday predictability, but no double-or-out rule beat buy-and-hold after costs and the reported data-snooping correction.

    Original publication ↗
    Methods & limitations
    Sample
    30 Dow Jones Industrial Average component stocks; sample dates require further verification.
    Frequency
    5-minute
    Patterns examined
    Candlestick timing rules and automated systems; see paper for definitions.
    Transaction costs
    Costs included in the reported comparisons
    Methods and benchmarks
    Compares timing rules to buy-and-hold; stepwise superior predictive ability test for the specified strategy family.
    Out-of-sample assessment
    Not established in this summary
    Multiple-testing treatment
    Stepwise superior predictive ability test for double-or-out strategies.
    Verification status
    Publisher abstract reviewed; exact sample dates and complete methods still require checking.
    Limitations
    Results are conditional on the market, intraday sampling, benchmarks, rules and modeled costs.
  3. 2014 · Stocks · Taiwan

    The profitability of candlestick charting in the Taiwan stock market

    Lu · Pacific-Basin Finance Journal

    Reported result: positive finding

    The study reported four profitable signals after its assumed transaction costs.

    Original publication ↗
    Methods & limitations
    Sample
    Taiwan stocks, 1992–2009
    Frequency
    Daily
    Patterns examined
    Twelve single-candle categories based on four-price-level definitions.
    Transaction costs
    Transaction costs included; assumptions require full-text review
    Methods and benchmarks
    Bootstrap analysis, subsample analysis and an out-of-sample exercise were reported.
    Out-of-sample assessment
    An out-of-sample exercise is reported; its exact design needs full-text verification.
    Multiple-testing treatment
    Bootstrap discussed; the scope of any multiple-testing adjustment needs further verification.
    Verification status
    Publisher abstract and indexed methodology reviewed; full results and assumptions require further review.
    Limitations
    Market-specific historical results do not establish profitability in other markets or periods.
  4. 2015 · Stocks · United States

    Trend definition or holding strategy: What determines the profitability of candlestick charting?

    Lu, Chen & Hsu · Journal of Banking & Finance

    Reported result: mixed findings

    Eight patterns were profitable with one holding strategy at an assumed 0.5% cost after the reported data-snooping adjustment; another holding strategy did not produce profits.

    Original publication ↗
    Methods & limitations
    Sample
    Dow Jones Industrial Average stocks; sample dates require full-text verification.
    Frequency
    Daily
    Patterns examined
    Three-day reversal patterns; multiple definitions of trend and holding strategies.
    Transaction costs
    Assumed 0.5% transaction cost; lower-cost sensitivity also reported
    Methods and benchmarks
    Compares trend definitions and holding strategies; includes a data-snooping adjustment.
    Out-of-sample assessment
    Not established in this summary
    Multiple-testing treatment
    The authors report adjustment for data snooping.
    Verification status
    Publisher abstract and institutional working-paper version reviewed; further full-text extraction is needed.
    Limitations
    The particular holding, execution and cost assumptions matter; results are not a universal property of a pattern.
  5. 2017 · Stocks · Thailand

    Profitability of candlestick charting patterns in the Stock Exchange of Thailand

    Tharavanij, Siraprapasiri & Rajchamaha · SAGE Open

    Reported result: limited evidence

    Mean returns for most patterns were not statistically different from zero; the tested filters generally did not improve results.

    Original publication ↗
    Methods & limitations
    Sample
    SET50 component stocks, July 2006–June 2016
    Frequency
    Daily
    Patterns examined
    Bullish and bearish reversals, with and without selected technical filters.
    Transaction costs
    Cost treatment needs further verification
    Methods and benchmarks
    Holding periods of 1, 3, 5 and 10 days; skewness-adjusted t tests and binomial tests.
    Out-of-sample assessment
    Not established in this summary
    Multiple-testing treatment
    Multiple-testing treatment not established in this summary.
    Verification status
    Full-text methods and results reviewed for the summarized findings.
    Limitations
    The study used a later SET50 constituent list; its sample selection and cost treatment deserve attention. Some individual results were significant and had substantial variability.
  6. 2020 · Forex · 24 currency pairs

    The bullish and the bearish engulfing patterns: beating the forex market or being beaten?

    Alanazi · The European Journal of Finance

    Reported result: positive finding

    The author reports profitable results for the tested approach, especially among major pairs, and stresses transaction costs.

    Original publication ↗
    Methods & limitations
    Sample
    24 currency pairs; observations from 2000–2018
    Frequency
    Daily and four-hour
    Patterns examined
    Bullish and bearish engulfing patterns with specified trading rules.
    Transaction costs
    Transaction costs examined; assumptions require fuller review
    Methods and benchmarks
    Daily candles were described as in-sample; four-hour candles as out-of-sample.
    Out-of-sample assessment
    Authors describe an out-of-sample four-hour exercise; it differs in frequency from the daily in-sample analysis.
    Multiple-testing treatment
    Multiple-testing treatment and timing of rule selection require fuller review.
    Verification status
    Publisher abstract and available methodological notes reviewed; complete methods and results not yet independently extracted.
    Limitations
    A different data frequency complicates comparisons; the result does not validate every engulfing strategy or currency pair.
  7. 2026 · Crypto · Multiple exchanges

    Intraday price forecasts using candlestick patterns in cryptocurrency markets

    Moser & Brauneis · International Review of Economics & Finance

    Reported result: positive finding

    Some patterns were followed by statistically significant returns in the expected direction under the study’s tests.

    Original publication ↗
    Methods & limitations
    Sample
    Nearly 400 cryptocurrencies and roughly 2,000 pairs on 36 exchanges, July 2018–January 2022
    Frequency
    Hourly
    Patterns examined
    55 reversal patterns detected using TA-Lib.
    Transaction costs
    Core return comparison excludes trading costs
    Methods and benchmarks
    Stepwise superior predictive ability test and robustness checks; close-to-close returns over multiple horizons.
    Out-of-sample assessment
    A genuinely untouched out-of-sample test is not established in this summary.
    Multiple-testing treatment
    Stepwise superior predictive ability test used to address data snooping.
    Verification status
    Publisher abstract and accessible paper text reviewed for data, forecasts and transaction-cost limitation.
    Limitations
    Entry at the pattern-closing price is assumed; results exclude costs in the core comparison and cannot be generalized to forex or equities.
Scroll the study list to see all seven records. Open an entry to inspect its methods.
How to interpret these findings

Result labels are navigation aids, not quality ratings. A “positive finding” means the authors reported at least one favorable result under their own test conditions. It does not establish consistent live profitability. “Mixed findings” and “limited evidence” summarize the particular results described, not the state of all worldwide research.

Markets, holding periods, price data, pattern rules, transaction costs, statistical adjustments, and truly untouched test samples all affect what a study can establish. In this collection, some methodology details remain marked as not fully verified. Those qualifications are intentional.

Reviewed against available publishers’ records on 16 November 2026. This is a curated teaching reference, not a systematic review, meta-analysis or independent replication.

FINANSIFIEDResearch notes · 7 selected publications

Why Do Candlestick Studies Reach Different Conclusions?

One study reports that a candlestick strategy was profitable. Another finds little evidence of an advantage. At first glance, those conclusions may look contradictory. But before deciding which study is right, we need to ask whether the researchers tested the same thing. A pattern’s name is only the beginning of a research question. Its definition, the market studied and the way a hypothetical trade is executed can all affect the result.

The Same Pattern Can Be Defined Differently

Consider a hammer. Most descriptions call for a small body and a long lower wick—but how small is small, and how long is long? One researcher might require the lower wick to be at least twice the length of the body. Another might compare the wick with the candle’s entire range. They may also use different rules to decide whether a decline preceded the formation.

Those choices determine which candles enter the study. Two researchers can both say they tested “hammers” while examining different sets of observations. This is why an objective definition matters: another researcher should be able to apply the same rules and identify the same formations in the same data.

Markets and Conditions Differ

A pattern studied in Taiwanese stocks during the 1990s and 2000s is not being tested under the same conditions as a pattern studied in cryptocurrency markets two decades later.

Markets differ in their participants, trading hours, liquidity, costs and access to information. Conditions can also change within the same market. A strategy examined during a prolonged rise may behave differently during a sharp decline or a quieter, sideways period.

A positive result in one setting is therefore a reason to investigate further. It is not a guarantee that the relationship will hold elsewhere—or persist in the future.

A Pattern Is Only One Part of a Strategy

Suppose two traders agree that a bullish engulfing pattern has appeared. They could still make very different decisions:

  • One buys immediately after the pattern completes; the other waits for confirmation.

  • One exits after the next candle; the other holds for several days.

  • One uses a stop-loss; the other exits only after a fixed period.

These are different strategies, even though they begin with the same pattern.

As we saw in the previous section, Lu, Chen and Hsu’s study of Dow stocks reported profitable results for some tested formations under one holding approach but not another. The finding illustrates why the entry pattern and the complete trading rule need to be evaluated separately.

Costs Can Turn a Small Advantage Into a Loss

Suppose a hypothetical strategy produces an average gain of 3 pips per trade before costs. If its average total trading costs are 4 pips per trade, its average result after those assumed costs is a 1-pip loss. The numbers are illustrative, but the mechanism is important. Depending on the market and strategy, costs may include the spread, commissions, financing charges and slippage. These frictions can be especially important for short-term strategies such as scalping, where the intended price move may be small. 

Execution assumptions matter, too. A backtest that enters at the exact price where a completed candle closes may obtain a price that would be difficult to capture after detecting the pattern and submitting an order. A result that looks attractive before costs is therefore a starting point. The next question is whether it remains attractive under realistic assumptions.

Researchers May Use Different Benchmarks

A pattern can be followed by a positive average return without adding useful information. For example, if stocks rose on average during a particular period, a bullish pattern might also be followed by gains simply because it occurred in a generally rising market. Researchers need a comparison that helps determine whether the pattern added anything beyond what might have happened anyway.

Depending on the question, that comparison might be the market’s average return, a matched set of periods without the pattern or a buy-and-hold strategy. Different benchmarks answer different questions—and can produce different conclusions.

Finding a Result Is Easier Than Confirming It

Researchers may try many patterns, markets, timeframes and exit rules. Some combinations can look successful by chance, especially if attention focuses on the strongest result after the search is complete. A stronger test considers the choices made during the search and evaluates the selected rules on data that were not used to develop them. Researchers may also use statistical methods designed to account for testing multiple possibilities.

Even then, one successful test is not the end of the investigation. Results are more persuasive when they can be reproduced and examined in other periods or markets—with the differences in those settings clearly stated. The reverse is also important: a study that does not find a statistically significant effect has not necessarily proved that the true effect is exactly zero. The test may have had limited data or substantial uncertainty. We need to look at the size and precision of the estimates, not just whether a result crosses a significance threshold.

Read the Result Alongside the Rules

When studies disagree, the explanation may be a difference in markets or methods. Sometimes a result may reflect chance or a weakness in the research. In other cases, a relationship may exist under one set of conditions but not another.

That is why the most useful research question is specific. Instead of asking whether a hammer works, ask whether a clearly defined hammer, in a specified market, followed by a specified decision, produced a result that held up against an appropriate benchmark after realistic costs.

The next section turns those questions into a practical checklist you can use when evaluating a candlestick claim.

How to Evaluate a Candlestick Strategy

Imagine someone claims that a bullish engulfing pattern is a reliable buy signal. They show you several charts where the pattern appeared just before a rally. The examples may be accurate, but they leave the most important questions unanswered: How were the examples selected? What happened after the other engulfing patterns? And could the results have been traded after costs? Use the following checklist to examine the claim.

Is the Pattern Defined Clearly?

“Look for a strong bullish candle” leaves considerable room for interpretation. A useful definition specifies the conditions another person would need to identify the same formation.

For an engulfing pattern, that might include:

  • Whether the second candle must engulf the first candle’s body or its entire range.

  • Whether the candle bodies must be different colors.

  • How the preceding trend is defined.

  • Which timeframe, session boundary and price source are used.

The definition does not need to be the only acceptable one. It does need to be clear enough to test and reproduce.

What Exactly Is Being Predicted?

“Price will go up” is incomplete. Does the claim concern the next candle, the next trading day or the next month? Is the forecast about direction, the size of the move or the probability of reaching a particular price?

A testable claim might be:

After a precisely defined bullish engulfing pattern appears on a daily chart, is the average return over the following five trading days different from an appropriate comparison?

That is a forecasting question. It can be studied without assuming that a trade is placed.

If It Is a Trading Strategy, Are All the Rules Specified?

A pattern alone does not tell us how a trade begins or ends. A complete strategy needs rules for:

  • Entry, when the position is opened and at what available price.

  • Exit, when the position is closed, including any stop-loss or profit target.

  • Position size, how much exposure is taken.

  • Filters, any additional conditions required before entering.

  • Costs, spreads, commissions, financing and execution assumptions.

If the pattern is only known when a candle closes, a test also needs to explain how the order could have been placed at the assumed entry price.

What Is the Result Being Compared With?

A strategy’s return is more informative when we know what it is being compared with. Depending on the question, a suitable benchmark might be the average return over similar periods, a comparable set of observations without the pattern or a passive investment in the same market. The comparison should reflect the claim: a test of short-term directional prediction and a test of long-term portfolio performance do not necessarily need the same benchmark. Ask whether the pattern added useful information beyond the market conditions in which it appeared.

How Much Evidence Is There?

Five successful chart examples do not tell us much about a pattern’s typical behavior. Look for the total number of occurrences, including unsuccessful and inconclusive cases. Also consider the length and diversity of the sample. A test covering only one short period may say little about how a strategy behaves under different conditions. A larger sample can help, but the raw number of trades is not the whole story: trades clustered around the same event or in overlapping periods may not provide as much independent information as their count suggests. Where possible, look for an estimate of uncertainty around the result—not just a single percentage.

Does the Result Survive Costs and Risk?

A high win rate can be appealing, but it does not tell us how much the strategy makes or loses. Suppose an illustrative strategy wins 60% of the time, gaining $50 on an average winning trade. It loses $100 on an average losing trade. Over 100 trades, that would be:

  • 60 wins × $50 = $3,000 gained

  • 40 losses × $100 = $4,000 lost

The result is a $1,000 loss before costs, despite the 60% win rate. Alongside win rate, look at average gains and losses, results after costs, drawdowns and the amount of risk taken. A strategy’s returns also need to be considered relative to a relevant benchmark and time period.

Was the Strategy Tested on Data It Had Not Already Seen?

A strategy can be adjusted repeatedly until it fits the historical period used to design it. This is called overfitting. A more informative process separates the data used to develop a rule from data reserved to evaluate it. In a walk-forward test, the researcher develops or updates the strategy using information available up to a point in time, then evaluates it over a later period. The process can be repeated through time.

Neither approach guarantees that a strategy will work in the future. If a researcher repeatedly checks the supposedly reserved period and adjusts the rules in response, that period is no longer a clean test of the final strategy.

It also helps to know how many patterns, filters and variations were tried before the reported result was selected.

Can Someone Else Reproduce the Result?

A credible claim should provide enough information for another person to understand—and ideally repeat—the analysis. That includes the data source, instruments, dates, pattern definitions, entry and exit rules, benchmark and cost assumptions. Independent replication in a different period or market can add useful evidence, although differences in data and methods must be considered when comparing results. A missing detail does not prove the strategy fails. It does mean there is less information available to assess the claim.

The Practical Question

A chart example can show what a pattern looks like. A backtest can help investigate what would have happened under specified historical assumptions. Neither, by itself, guarantees future performance.

The goal of this checklist is to move from “I found a pattern before a big move” to a more useful question:

After defining the pattern and trading rules in advance, is there evidence that the result adds value beyond a suitable benchmark, survives realistic costs and holds up on new data?

Why Are Candlestick Patterns So Appealing to Retail Traders?

A candlestick chart can make a complicated market feel more approachable. Instead of beginning with a table of prices or a statistical model, a new trader can look at a shape, learn its name and understand the price information it summarizes. That is a useful starting point. The difficulty comes when learning to recognize the shape begins to feel like evidence that the next move is predictable.

They Turn Price Data Into Memorable Shapes

A hammer, a shooting star and an evening star are easier to remember than a set of rules about opening prices, closing prices and the relative lengths of wicks. The visual format also lets readers compare formations quickly. That is one reason candlestick guides often use illustrated catalogues and cheat sheets: they give beginners a vocabulary for describing what they see. There is nothing inherently wrong with that. A well-designed visual can make unfamiliar information easier to understand. But knowing the name of a pattern is different from knowing whether it improves a forecast.

They Offer a Story About the Market

Pattern descriptions often turn price movements into a story about buyers and sellers. A long lower wick might be described as sellers pushing the price down before buyers drive it back up. An engulfing candle might be presented as one side taking control from the other. These narratives are intuitive. They give the shape a meaning that is easier to remember than four price values. The caution is that a candle does not reveal the identity, motives or full sequence of actions of market participants. A story may be plausible, but it needs additional evidence before it can be treated as an explanation of what caused the move.

Historical Examples Make the Lesson Feel Clear

An educator can show a hammer followed by a rally or a shooting star followed by a decline. A reader can immediately see the formation and its apparent outcome. Examples like these are useful for teaching what a pattern looks like. They are less useful for establishing how well it predicts unless we also know how often the same pattern appeared without the expected move. This is a subtle difference. An example answers “What does this formation look like?” A properly designed test asks “What tended to happen after all formations that met the same definition?”

They Fit Naturally Into Trading Education

Candlestick patterns can be organized into short lessons, lists, diagrams and printable references. A beginner can learn one formation, find it on a chart and feel they have made progress. That sense of progress is understandable: identifying a pattern is a skill that can be practiced. But it is only one part of evaluating a trading idea. A complete strategy also requires entry and exit rules, risk controls and evidence that its results are not explained by the market’s general behavior, chance or unrealistic assumptions.

Educational providers do not all make the same claims. Some present patterns as possible signals and emphasize context; others use more confident language. For example, BabyPips’ candlestick reference describes potential interpretations, while also noting that price can move in the opposite direction. IG’s candlestick guide discusses patterns in relation to additional context such as support, resistance and indicators.

Those added conditions may be worth investigating, but they introduce another question: has the combination itself been tested? As we saw earlier, adding a filter does not automatically improve a strategy’s results.

The Difference Between Learning and Having an Edge

Candlesticks can provide a useful visual language. They make it possible to describe where a price opened, how far it moved and where it closed. They can also help traders formulate ideas worth investigating. But financial markets do not reward a trader simply for recognizing a familiar shape. Whether that recognition improves a decision depends on the evidence behind the specific claim—and, for a trading strategy, what remains after costs and risk are considered.

So, if memorizing patterns is not enough to establish an advantage, what can candlesticks actually be useful for?

What Can Candlesticks Be Useful For?

After all these questions about definitions, prediction and profitability, it is reasonable to ask whether candlestick charts are worth using at all. They can be. A candlestick is a way of organizing price information, and that information can be useful even when we have not established that a particular formation predicts the next move.

Seeing the Price Range at a Glance

A candle brings the open, high, low and close together in one visual. At a glance, a reader can see whether the period ended above or below its opening price and how far the recorded price moved beyond those two points. For example, two assets might both open at $100 and close at $102. If one traded between $99 and $103 while the other ranged from $92 to $108, their candles would show that important difference.

The candles summarize the ranges; they do not explain what caused them.

Comparing Periods and Describing Events

Candles also make it easier to compare one period with another. A trader or analyst can use them to describe a wider range around an announcement, a gap between sessions or a move that reversed before the period ended. They provide a shared visual language. Saying that a candle has a long upper wick is a concise way to describe a high well above its body. Other readers can then inspect the same chart and understand the observation—provided they know the timeframe and price source.

Generating Questions Worth Testing

A recognizable formation can also be the starting point for research. A trader might notice that certain long-wicked candles appear around periods of high volatility and ask:

Do returns after these candles differ from returns after comparable periods without them?

That question does not assume the answer. It can be turned into a test by defining the candle, selecting an appropriate comparison and examining the results across data that were not used to invent the rule.Some research has found associations between defined patterns and later price movements under particular conditions. Other studies have found little value in the strategies they tested. The task is to evaluate the specific evidence rather than assume that every formation has—or lacks—predictive power.

Putting Candles in Context

Candlesticks can be used alongside other information, including market conditions, trading costs and relevant news. Additional information may help a reader frame a better question, but it does not automatically make a pattern more reliable.

If a strategy combines a candle with a trend rule, an indicator or an economic event, that combination becomes the idea to evaluate. It should be judged on its own evidence.

A candlestick is therefore useful for what it directly shows: a summary of recorded prices over a chosen interval. It can also help generate and communicate hypotheses. What it cannot do on its own is prove why the price moved or establish that the next move is predictable.

Conclusion: Recognizing a Pattern Is Not the Same as Having an Edge

Candlestick patterns offer a visual language for describing price movements. Once you understand the body and wicks, you can identify formations, compare periods and communicate what happened on a chart. But the name attached to a formation does not tell you what will happen next. A hammer can appear before a rally or before a further decline. A recognizable shape can emerge from a sequence with no predictive signal. And even when research finds a relationship between a pattern and subsequent returns, turning that relationship into a profitable strategy requires further evidence.

The research we examined offers reasons to remain open-minded as well as skeptical. Some studies report positive findings under specific conditions; others find little value in the patterns or strategies they tested. The market, definition, timeframe, exit rules and costs all matter. You do not have to abandon candlestick charts to take those limitations seriously. Use them to describe what the recorded price did, explore the ideas traders attach to formations and ask better questions about the market. Then take the most important step: test the claim, not just your ability to recognize the pattern.

References20 sources · 3 research areas

Selected academic, institutional and technical sources supporting the discussion of candlestick patterns, market structure and research methods.

01 / Candlestick Pattern Research

  1. Marshall, B. R., Young, M. R., & Rose, L. C. (2006).

    Candlestick Technical Trading Strategies: Can They Create Value for Investors?

    Journal of Banking & Finance, 30(8), 2303–2323.

  2. Duvinage, M., Mazza, P., & Petitjean, M. (2013).

    The Intra-Day Performance of Market Timing Strategies and Trading Systems Based on Japanese Candlesticks

    Quantitative Finance, 13(7), 1059–1070.

  3. Lu, T.-H. (2014).

    The Profitability of Candlestick Charting in the Taiwan Stock Market

    Pacific-Basin Finance Journal, 26, 65–78.

  4. Lu, T.-H., Chen, Y.-C., & Hsu, Y.-C. (2015).

    Trend Definition or Holding Strategy: What Determines the Profitability of Candlestick Charting?

    Journal of Banking & Finance, 61, 172–183.

  5. Tharavanij, P., Siraprapasiri, V., & Rajchamaha, K. (2017).

    Profitability of Candlestick Charting Patterns in the Stock Exchange of Thailand

    SAGE Open, 7(4).

  6. Alanazi, A. S. (2020).

    The Bullish and the Bearish Engulfing Patterns: Beating the Forex Market or Being Beaten?

    The European Journal of Finance, 26(15), 1484–1505.

  7. Moser, S., & Brauneis, A. (2026).

    Intraday Price Forecasts Using Candlestick Patterns in Cryptocurrency Markets

    International Review of Economics & Finance, 108, 105158.

02 / Market Structure, Pattern Perception and Research Methods

  1. Evans, M. D. D., & Lyons, R. K. (2002).

    Order Flow and Exchange Rate Dynamics

    Journal of Political Economy, 110(1), 170–180.

  2. Nickerson, R. S. (2002).

    The Production and Perception of Randomness

    Psychological Review, 109(2), 330–357.

  3. Lo, A. W., & MacKinlay, A. C. (1990).

    Data-Snooping Biases in Tests of Financial Asset Pricing Models

    The Review of Financial Studies, 3(3), 431–467.

  4. CFA Institute.

    Backtesting & Simulation

    2026 CFA Program refresher reading.

  5. CFA Institute.

    Active Equity Investing: Strategies

    2026 CFA Program refresher reading.

  6. Bank for International Settlements (2019).

    FX Trade Execution: Complex and Highly Fragmented

    BIS Quarterly Review.

03 / Chart Construction, Pattern Definitions and Data

  1. OANDA.

    User Guide for OANDA Web Platform

    Platform chart configuration and timeframes.

  2. OANDA.

    Using Graph Features

    Currency-chart construction and settings.

  3. TA-Lib.

    Candlestick Settings

    Pattern-recognition thresholds and configurable settings.

  4. TA-Lib.

    Functions: Pattern Recognition

    Technical-analysis function reference.

  5. Board of Governors of the Federal Reserve System.

    Foreign Exchange Rates—H.10: Historical Rates for the Euro

    Data source for the historical pattern-perception exercise.

  6. BabyPips.

    Candlestick Patterns Cheat Sheet

    Conventional pattern descriptions and classifications.

  7. IG.

    16 Candlestick Patterns Every Trader Should Know

    Example of educational presentations of traditional candlestick interpretations.

FINANSIFIED · SHARE

Share this article

Pass this research along, wherever the conversation is happening.

To share this page, copy its address from your browser. Social sharing buttons activate when JavaScript is enabled.