Based on Gregory L. Morris, Candlestick Charting Explained: Timeless Techniques for Trading Stocks and Futures (3rd Edition, 2006)
Gregory Morris's Candlestick Charting Explained has been the definitive English- language reference on Japanese candlestick patterns since its first edition in 1992. The third edition, published in 2006, represents a significant update that incorporates over a decade of additional research, backtesting data, and practical application experience.
While the first and second editions established the pattern taxonomy and basic interpretation framework, the third edition adds several critical dimensions: rigorous statistical testing of pattern reliability, improved filtering techniques to reduce false signals, integration with Western technical analysis tools, and practical guidance for algorithmic implementation.
The Collector's Edition (covered in a separate summary) focused on the comprehensive catalog of all candlestick patterns with traditional interpretations. This third edition summary focuses specifically on what is new or refined: the empirical testing results, the filtering methodology, and the modern integration techniques that move candlestick analysis from art to applied science.
By the third edition, Morris's approach had matured significantly. He moved from a pattern-catalog approach ("learn all the patterns") to a systematic methodology ("test the patterns, filter the signals, confirm with other tools"). This shift reflects the broader maturation of technical analysis from visual pattern recognition to quantitative, evidence-based trading.
| Feature | 1st/2nd Editions | 3rd Edition |
|---|---|---|
| Pattern count | Comprehensive catalog | Same catalog with statistical annotations |
| Backtesting | Minimal or anecdotal | Systematic backtesting with reliability percentages |
| Filtering | Basic (prior trend, confirmation) | Multi-layer filtering (trend, indicators, volume) |
| Modern tools | Brief mentions | Detailed integration with RSI, MACD, Bollinger Bands |
| Algorithmic notes | None | Pattern recognition programming guidance |
| Market application | Primarily equities/futures | Extended to forex, ETFs, options |
| Timeframes | Daily focus | Multi-timeframe analysis framework |
The third edition incorporates data through the mid-2000s, which means it covers several additional market cycles not available in the earlier editions:
This additional data provides more robust testing across both bull and bear market conditions, which is essential for validating reversal patterns.
Morris increasingly emphasizes that candlestick patterns should never be used in isolation. The third edition makes this a central theme: candlestick patterns are components of a trading system, not a complete system in themselves. This represents a significant shift from the first edition, which, like most early candlestick books, tended to present patterns as standalone signals.
The third edition provides more precise mathematical definitions for pattern identification, replacing the somewhat subjective criteria of earlier editions.
Doji Refinement:
PREVIOUS DEFINITION:
Open equals Close (or very nearly so)
3RD EDITION DEFINITION:
|Close - Open| / (High - Low) < threshold
Where threshold varies by context:
Short-term trading: 0.05 (5% of range)
Swing trading: 0.10 (10% of range)
Long-term analysis: 0.15 (15% of range)
Additional refinement by doji type:
Standard Doji: Short upper and lower shadows
Long-Legged Doji: Upper and lower shadows > 2x average range
Dragonfly Doji: Lower shadow > 3x body, minimal upper shadow
Gravestone Doji: Upper shadow > 3x body, minimal lower shadow
Hammer/Hanging Man Refinement:
PREVIOUS DEFINITION:
Small body at upper end, long lower shadow
3RD EDITION CRITERIA:
1. Lower shadow ≥ 2x the body length
2. Upper shadow ≤ 10% of total range (ideally zero)
3. Body is in the upper 1/3 of the total range
4. Color of body is secondary but provides context:
- White/green body = slightly more bullish for hammer
- Black/red body = slightly more bearish for hanging man
5. MUST appear after a definable trend:
- Hammer: after a downtrend of at least 5 periods
- Hanging Man: after an uptrend of at least 5 periods
BULLISH ENGULFING (refined):
1. First candle is bearish (close < open)
2. Second candle is bullish (close > open)
3. Second candle's body COMPLETELY engulfs first candle's body
(second open < first close AND second close > first open)
4. Must appear in a downtrend (at least 5 periods declining)
5. ENHANCEMENT (3rd ed): Volume on second candle should exceed
volume on first candle by at least 50%
6. ENHANCEMENT (3rd ed): Second candle's body should be at least
1.5x the average body size of the prior 10 candles
BEARISH ENGULFING: Mirror image with analogous volume/size requirements
Morris tightens the definition of the gap requirements, noting that in modern 24-hour electronic markets, traditional gaps are less common. He allows for "quasi-gaps" where the second candle's body does not overlap with the first candle's body, even if the shadows do overlap.
Morris tested candlestick patterns across a broad universe of securities using the following methodology:
| Pattern | Type | 5-Day Accuracy | 10-Day Accuracy | Reliability Rating |
|---|---|---|---|---|
| Bullish Engulfing | Bullish reversal | 53-58% | 55-62% | High |
| Bearish Engulfing | Bearish reversal | 52-57% | 54-60% | High |
| Morning Star | Bullish reversal | 55-62% | 58-65% | Very High |
| Evening Star | Bearish reversal | 54-60% | 57-63% | Very High |
| Hammer | Bullish reversal | 50-55% | 53-59% | Moderate |
| Hanging Man | Bearish reversal | 48-53% | 50-55% | Low-Moderate |
| Doji Star | Reversal (context) | 49-54% | 52-58% | Moderate |
| Piercing Line | Bullish reversal | 51-56% | 54-60% | Moderate-High |
| Dark Cloud Cover | Bearish reversal | 50-55% | 53-58% | Moderate |
| Three White Soldiers | Bullish | 56-63% | 59-66% | Very High |
| Three Black Crows | Bearish | 55-61% | 57-64% | Very High |
Finding 1: Multi-candle patterns are more reliable than single-candle patterns. Three-candle patterns (Morning Star, Evening Star, Three White Soldiers) consistently outperformed single-candle patterns (Hammer, Doji). The additional candles provide more information and more confirmation.
Finding 2: Longer measurement windows improve accuracy. A pattern that shows 53% accuracy at 5 days might show 60% at 10 days and 62% at 20 days. This suggests candlestick patterns are better suited for swing trading (5-20 days) than for day trading.
Finding 3: Pattern filtering dramatically improves reliability. Raw pattern recognition typically shows 50-60% accuracy. Adding trend filters raises this to 55-65%. Adding volume confirmation raises it to 60-70%. Adding indicator confirmation can push accuracy to 65-75%.
Finding 4: Bearish patterns are slightly less reliable than bullish ones. This is likely because the long-term upward bias of equity markets reduces the effectiveness of bearish reversal signals.
Morris notes that continuation patterns (windows/gaps, three methods) are generally less reliable as standalone signals than reversal patterns. They work best as confirmation of an already-identified trend rather than as primary entry signals.
This is the most practically valuable addition in the third edition.
Morris advocates a three-layer approach to candlestick signal validation:
LAYER 1: TREND CONTEXT
- Is there a definable prior trend for reversal patterns?
- Minimum trend duration: 5 periods (Morris prefers 8-10)
- Trend identification method: Moving average slope, swing highs/lows,
or linear regression
LAYER 2: CANDLESTICK PATTERN
- Does the pattern meet all precise identification criteria?
- Is the pattern "clean" (close to ideal form) or "messy"?
- Rating: Strong, Moderate, Weak
LAYER 3: CONFIRMATION
- Does at least one confirming indicator support the signal?
- Volume confirmation
- Oscillator confirmation (RSI, Stochastic)
- Support/resistance level alignment
- Moving average proximity
TRADE SIGNAL:
All three layers aligned → HIGH CONFIDENCE signal
Two layers aligned → MODERATE CONFIDENCE signal
One layer only → LOW CONFIDENCE — do not trade
Morris emphasizes the importance of waiting for a confirmation candle — the candle that follows the pattern — before acting. For bullish patterns, the confirmation candle should open higher and close higher. For bearish patterns, it should open lower and close lower.
CONFIRMATION RULES:
Bullish pattern confirmation:
Next candle opens above midpoint of signal candle's body
AND closes above signal candle's high
→ CONFIRMED
Bullish pattern weak confirmation:
Next candle closes above signal candle's close
but below signal candle's high
→ PARTIALLY CONFIRMED (reduce position size)
Bullish pattern failure:
Next candle closes below signal candle's low
→ PATTERN FAILED (no trade)
The third edition places far greater emphasis on volume than previous editions:
Morris provides specific integration rules:
MOVING AVERAGE INTEGRATION:
Bullish candlestick pattern AT the 50-day or 200-day moving average:
→ Signal reliability increases significantly
→ The MA acts as a support level that validates the reversal
Example:
Stock in uptrend pulls back to 50-day MA
Hammer or Bullish Engulfing forms at the MA
→ High-probability long entry
Bearish candlestick pattern AT a declining moving average:
→ Resistance from the MA confirms the bearish signal
MOVING AVERAGE TREND FILTER:
Only take bullish patterns when price > 200-day MA (or MA is rising)
Only take bearish patterns when price < 200-day MA (or MA is falling)
This single filter eliminates many false signals.
RSI INTEGRATION:
Bullish candlestick + RSI < 30 (oversold):
→ High-probability reversal signal
→ Even better if RSI shows positive divergence
Bearish candlestick + RSI > 70 (overbought):
→ High-probability reversal signal
→ Even better if RSI shows negative divergence
RSI DIVERGENCE + CANDLESTICK:
Price makes new low, RSI makes higher low (positive divergence)
THEN bullish candlestick pattern forms
→ One of the highest-probability setups available
BOLLINGER BAND INTEGRATION:
Bullish reversal at lower Bollinger Band:
- Price touches or pierces the lower band
- Bullish candlestick pattern forms
- Price returns inside the bands
→ Strong buy signal (mean-reversion + reversal confirmation)
Bearish reversal at upper Bollinger Band:
- Price touches or pierces the upper band
- Bearish candlestick pattern forms
- Price returns inside the bands
→ Strong sell signal
SQUEEZE + CANDLESTICK:
Bollinger Bands narrow (volatility contraction)
Decisive candlestick breakout pattern
→ Potential trend initiation signal
Morris notes that MACD is most useful for confirming the trend direction rather than timing specific entries. A bullish candlestick pattern with MACD above zero (or crossing above its signal line) has higher reliability than one with MACD below zero and falling.
Morris expands the discussion of the Abandoned Baby pattern, which he considers one of the most powerful reversal signals when it appears clearly:
BULLISH ABANDONED BABY:
Candle 1: Long bearish candle in a downtrend
Candle 2: Doji that gaps down from Candle 1
(Candle 2's HIGH is below Candle 1's LOW)
Candle 3: Bullish candle that gaps up from the doji
(Candle 3's LOW is above Candle 2's HIGH)
The doji is completely isolated — "abandoned" — by gaps on both sides.
This is extremely rare but extremely reliable (>75% in testing).
Note: True gaps are rare in modern markets with extended hours.
Morris allows modified criteria where shadows may overlap slightly.
A two-candle pattern consisting of a marubozu followed by an opposite-colored marubozu with a gap between them. Morris's testing found this to be one of the most reliable patterns statistically, though it is relatively rare.
Morris provides expanded treatment of these less common patterns with new backtesting data, noting that while individually less reliable, they become meaningful when combined with support/resistance levels.
Updated treatment with specific criteria for distinguishing between significant Belt Holds and normal long-body candles. The key criterion: the candle must open at or very near its extreme (high for bearish, low for bullish) with no shadow at the opening end.
The third edition introduces a formal reliability ranking that categorizes all patterns into tiers:
TIER 1 — MOST RELIABLE (>60% accuracy with proper filtering):
- Morning Star / Evening Star
- Three White Soldiers / Three Black Crows
- Bullish/Bearish Engulfing (with volume confirmation)
- Abandoned Baby
- Kicking Pattern
- Three Inside Up / Three Inside Down
TIER 2 — RELIABLE (55-60% accuracy with proper filtering):
- Piercing Line / Dark Cloud Cover
- Hammer (with confirmation candle)
- Bullish/Bearish Harami (with indicator confirmation)
- Rising/Falling Three Methods
- Upside/Downside Gap Two Crows
TIER 3 — MODERATE (50-55% accuracy, requires strong additional confirmation):
- Single Doji patterns
- Hanging Man
- Shooting Star
- Belt Hold
- Homing Pigeon
- Matching Low/High
TIER 4 — LOW RELIABILITY (not recommended as primary signals):
- Most rare/obscure patterns
- Patterns without trend context
- Any pattern against the major trend
Morris introduces the concept of "context multipliers" — factors that increase or decrease a pattern's base reliability:
| Context Factor | Multiplier Effect |
|---|---|
| Pattern at major support/resistance | +10-15% accuracy |
| Volume confirmation (>150% of average) | +5-10% accuracy |
| RSI/Stochastic extreme alignment | +5-10% accuracy |
| Moving average convergence | +5-8% accuracy |
| Pattern against the major trend | -10-15% accuracy |
| Low volume on signal candle | -5-10% accuracy |
| Choppy/range-bound market | -5-10% accuracy |
| Multiple indicators conflicting | -10-15% accuracy |
The third edition introduces a multi-timeframe framework that was absent from earlier editions:
MULTI-TIMEFRAME HIERARCHY:
Weekly chart: Determine the major trend direction
Daily chart: Identify candlestick reversal/continuation patterns
Intraday: Fine-tune entry and exit timing
RULE: Only trade candlestick patterns on the daily chart that are
aligned with the weekly trend direction.
Example:
Weekly: Uptrend (price above rising 40-week MA)
Daily: Pullback produces bullish hammer at support
→ Take the bullish signal (aligned with weekly trend)
Weekly: Downtrend (price below falling 40-week MA)
Daily: Pullback produces bullish hammer
→ Ignore or reduce confidence (fighting the major trend)
Morris explains how to combine multiple candles into a single "composite" candle for pattern simplification:
COMBINING TWO DAILY CANDLES INTO ONE:
Open = First candle's open
Close = Second candle's close
High = MAX(first high, second high)
Low = MIN(first low, second low)
APPLICATION:
A bullish engulfing pattern (two candles) combines into a single
hammer-like candle → confirms the bullish interpretation.
A morning star (three candles) combines into a single long white
candle with a lower shadow → confirms bullish reversal strength.
This technique helps validate pattern interpretation by checking
whether the "story" told by multiple candles is consistent when
simplified.
Morris provides specific volume expectations for each major pattern:
BULLISH ENGULFING VOLUME PROFILE:
Day 1 (bearish candle): Normal or below-average volume
Day 2 (bullish engulfing candle): Above-average volume (>1.5x avg)
→ Interpretation: Heavy buying overwhelms selling pressure
MORNING STAR VOLUME PROFILE:
Day 1 (long bearish candle): Above-average volume (selling climax)
Day 2 (small body/doji): Low volume (indecision, selling exhausted)
Day 3 (long bullish candle): Above-average volume (buyers return)
→ The "volume valley" on Day 2 is key
HAMMER VOLUME PROFILE:
Signal day: High volume, especially if the intraday pattern shows
volume concentrated in the second half of the session (recovery)
→ Heavy volume hammer is much more reliable than light volume
WARNING SIGNALS:
Bullish pattern with DECLINING volume:
→ Questionable — buyers not showing conviction
→ Reduce position size or wait for additional confirmation
Bearish pattern with DECLINING volume:
→ May still work (markets can fall on low volume)
→ But less reliable than bearish pattern with heavy volume
Volume spike WITHOUT a candlestick pattern:
→ Often precedes a pattern by 1-2 candles
→ Monitor closely for subsequent pattern formation
Morris acknowledges that algorithmic implementation requires decisions about thresholds that do not exist in visual pattern recognition. He provides guidance:
BODY SIZE THRESHOLDS:
"Long" body: Body > 1.5x the average body size of prior 10 candles
"Short" body: Body < 0.5x the average body size of prior 10 candles
"Doji": Body < 0.1x the total range (High - Low)
SHADOW THRESHOLDS:
"Long" shadow: Shadow > 2x the body length
"Short" shadow: Shadow < 0.3x the body length
"No" shadow: Shadow < 0.05x the total range
TREND DEFINITION:
Uptrend: Price has risen > 5% over the past 10 candles
OR 10-period MA is rising
OR higher highs and higher lows visible
Downtrend: Price has declined > 5% over the past 10 candles
OR 10-period MA is falling
OR lower highs and lower lows visible
In modern electronic markets, especially forex, traditional gaps are rare. Morris suggests adapting gap requirements:
"Candlestick patterns are a language of the market, but like any language, they must be understood in context. A word in isolation can have many meanings; a word in a sentence is clear."
"The single most important lesson I have learned in decades of working with candlestick charts is that the pattern is only the beginning of the analysis, not the end."
"Pattern filtering is not optional. Unfiltered candlestick signals are only marginally better than random. Properly filtered signals are a genuine edge."
"Volume is the fuel that drives the market. A candlestick pattern without volume confirmation is like a car without gasoline — it may look right, but it is not going anywhere."
"The trend is the single most important filter for any candlestick pattern. Reversal patterns must have something to reverse. Continuation patterns must have something to continue."
"Multi-candle patterns outperform single-candle patterns because they contain more information. The market is telling you a story over two or three days, not just one."
"The combination of a candlestick reversal pattern at a key support or resistance level, with above-average volume and a confirming oscillator reading, represents one of the highest-probability setups available to the technical trader."
"I have tested every candlestick pattern I know across multiple markets, timeframes, and market conditions. The results confirm that a handful of patterns, properly filtered, provide a genuine statistical edge. Most rare or exotic patterns do not."
"Do not memorize patterns for the sake of memorization. Learn the five or six most reliable patterns deeply, understand their logic, and develop the discipline to wait for properly confirmed signals."
"The confirmation candle is not an afterthought — it is an integral part of the signal. Acting before confirmation is speculation. Acting after confirmation is systematic trading."
The third edition of Candlestick Charting Explained represents the maturation of candlestick analysis from a pattern-catalog approach to an evidence-based, systematic methodology. Morris's key contribution is the multi-layer filtering framework: trend context, pattern recognition, and indicator confirmation working together to produce high-probability signals. For the modern practitioner, this edition provides the bridge between traditional Japanese candlestick wisdom and contemporary quantitative trading practice.