Based on Gregory Morris, Japanese Candlestick Charting Techniques (θ‘ηεΎη²Ύθ§£)
Candlestick charting traces its roots to 18th-century Japan, where Munehisa Homma β a legendary rice trader from Sakata β developed techniques to analyze the price of rice contracts at the Dojima Rice Exchange in Osaka. Homma's methods evolved into what we now recognize as candlestick analysis. He reportedly executed over 100 consecutive winning trades and amassed enormous wealth, earning him the title "God of the Markets."
The Japanese approach predates Western bar charting by more than a century. While Charles Dow was formulating his theories in the 1880s, Japanese traders had already spent 200 years refining visual price analysis. Steve Nison introduced these techniques to the Western world in the early 1990s; Gregory Morris's work builds on that foundation with rigorous statistical testing and practical implementation guidance.
The candlestick philosophy rests on several principles:
| <- Upper Shadow (Wick)
+-+
| | <- Real Body (Open to Close)
| | White/Green = Close > Open (bullish)
+-+ Black/Red = Close < Open (bearish)
| <- Lower Shadow (Tail)
Key measurements used throughout this specification:
body_size = abs(close - open)
upper_shadow = high - max(open, close)
lower_shadow = min(open, close) - low
total_range = high - low
body_midpoint = (open + close) / 2
Definition: A candle where the open and close are equal (or nearly equal).
Identification rules:
body_size <= total_range * 0.05 (body is less than 5% of total range)Variants:
| Variant | Upper Shadow | Lower Shadow | Meaning |
|---|---|---|---|
| Standard Doji | Roughly equal | Roughly equal | Pure indecision |
| Long-Legged Doji | Very long | Very long | Extreme indecision, often at turning points |
| Dragonfly Doji | Minimal | Very long | Bullish reversal at bottoms |
| Gravestone Doji | Very long | Minimal | Bearish reversal at tops |
| Four-Price Doji | None | None | Open = High = Low = Close; ultra-low volume |
Significance: A doji after a strong trend signals that the dominant side is losing conviction. It does not itself reverse the trend; it warns that the trend is vulnerable.
Definition: A small real body at the upper end of the range with a long lower shadow. Appears in a downtrend. The most potent single-candle bullish reversal signal.
Identification rules:
lower_shadow >= body_size * 2.0 (lower shadow at least twice the body)upper_shadow <= body_size * 0.3 (upper shadow very small or absent)Psychology: Sellers pushed price sharply lower during the session but buyers stepped in aggressively and drove price back near the open. The long lower tail is evidence of demand.
Definition: Identical shape to the hammer, but appears in an uptrend. Bearish warning.
Identification rules:
Critical distinction: The hanging man is far less reliable than the hammer without confirmation. Many hanging men fail in strong uptrends. Morris emphasizes that a hanging man without next-day confirmation should be ignored.
Definition: A small real body at the lower end of the range with a long upper shadow. Appears in an uptrend. Bearish reversal signal.
Identification rules:
upper_shadow >= body_size * 2.0lower_shadow <= body_size * 0.3Psychology: Buyers pushed price to new highs but sellers overwhelmed them, driving the close back near the open. The long upper shadow is a visual record of rejected higher prices.
Definition: Same shape as shooting star, but appears in a downtrend. Bullish signal.
Identification rules:
Definition: A candle with no shadows (or extremely small shadows). Represents total dominance by one side.
Identification rules:
upper_shadow <= total_range * 0.01lower_shadow <= total_range * 0.01body_size >= total_range * 0.95Variants:
Significance: A marubozu indicates extreme conviction. When it appears after a period of consolidation, it often marks the beginning of a sustained move.
Definition: A small real body with upper and lower shadows that are longer than the body.
Identification rules:
body_size <= total_range * 0.30upper_shadow > body_sizelower_shadow > body_sizeSignificance: Mild indecision. Less powerful than a doji but still signals waning momentum when it appears after a strong directional move.
Definition: A two-candle pattern where a large white body completely engulfs the prior small black body. Appears in a downtrend.
Identification rules:
open_2 < close_1 AND close_2 > open_1Definition: Mirror of bullish engulfing. A large black body engulfs the prior small white body in an uptrend.
Identification rules:
open_2 > close_1 AND close_2 < open_1Definition: The reverse of engulfing β a small body is contained within the prior large body. "Harami" means "pregnant" in Japanese.
Bullish Harami rules (in downtrend):
open_2 > close_1 AND close_2 < open_1 (for bearish Candle 1)Bearish Harami rules (in uptrend):
Harami Cross: When Candle 2 is a doji, the pattern is called a Harami Cross and carries significantly more reversal weight.
Definition: A bullish reversal pattern in a downtrend. The second candle opens below the prior low and closes above the midpoint of the prior body.
Identification rules:
open_2 < low_1 (gaps below prior low)close_2 > body_midpoint_1 (closes above midpoint of Candle 1's body)close_2 < open_1 (does not fully engulf β that would be an engulfing pattern)Definition: Bearish mirror of the piercing line. In an uptrend, the second candle opens above the prior high and closes below the midpoint of the prior body.
Identification rules:
open_2 > high_1close_2 < body_midpoint_1close_2 > close_1 (does not fully engulf)Definition: Two or more candles with matching highs (tweezer top) or matching lows (tweezer bottom).
Tweezer Bottom rules:
abs(low_1 - low_2) <= ATR * 0.01 (lows are approximately equal)Tweezer Top rules:
abs(high_1 - high_2) <= ATR * 0.01Significance: Tweezers mark precise support/resistance levels. Their reliability increases when the component candles are themselves recognizable patterns (e.g., a tweezer bottom formed by a hammer and a bullish engulfing).
Definition: A three-candle bullish reversal at the bottom of a downtrend.
Identification rules:
max(open_2, close_2) < close_1 (ideal gap; in stocks, gap is common; in forex, overlap is acceptable if Candle 2's body is very small)close_3 > body_midpoint_1Morning Doji Star: When Candle 2 is a doji, the pattern is called a Morning Doji Star and is considered more powerful.
Definition: Bearish mirror of the morning star. Three-candle reversal at the top.
Identification rules:
close_3 < body_midpoint_1Definition: Three consecutive long bullish candles, each opening within the prior body and closing at or near its high. Strong bullish continuation/reversal signal.
Identification rules:
open_n > close_(n-1) * 0.7 + open_(n-1) * 0.3 (approximate β opens in lower portion of prior body)upper_shadow_n <= body_size_n * 0.2Definition: Bearish mirror of three white soldiers. Three consecutive long bearish candles, each opening within the prior body and closing near its low.
Identification rules:
lower_shadow_n <= body_size_n * 0.2Definition: A bullish harami followed by a confirmation candle that closes above the first candle's open.
Identification rules:
close_3 > open_1Definition: A bearish harami followed by a confirmation candle that closes below the first candle's close.
Identification rules:
close_3 < close_1Definition: A bullish continuation pattern. A long white candle is followed by three (sometimes two or four) small declining candles that stay within the range of the first candle, then a final large white candle closes above the first candle's high.
Identification rules:
Definition: Bearish mirror. A long black candle, three small rising candles within its range, then a final large black candle closing below the first candle's low.
Definition: In an uptrend, a bullish candle gaps up from the prior bullish candle. The third candle is bearish and opens within the body of the second candle, closing within the gap but not filling it completely.
Identification rules:
low_2 > high_1)close_3 > high_1Definition: Bearish mirror. A gap-down sequence where the third candle partially fills the gap but does not close it.
Definition: In an uptrend, a bearish candle is followed by a bullish candle that opens at the same price as the bearish candle's open. The trend continues upward.
Identification rules (bullish):
abs(open_2 - open_1) <= ATR * 0.005Volume is the single most important confirmation tool for candlestick patterns.
Rules:
Morris stresses that many patterns require the next trading day to confirm:
The most critical rule in candlestick analysis:
A hammer in the middle of a sideways range is not a reversal signal. An evening star in a downtrend is meaningless. Trend context is non-negotiable.
Patterns occurring at established support or resistance levels carry 2-3x more weight:
Integration method:
Example: Price declines to the 200-day moving average, a well-known dynamic support. A hammer forms with its low precisely touching the 200-day MA. The next day opens higher. Entry: above the hammer's high. Stop: below the hammer's low (which is also below the 200-day MA). This combines candlestick logic with Western structural analysis.
| Situation | Signal Strength |
|---|---|
| Bullish pattern at rising 50-day MA | Strong |
| Bearish pattern at declining 50-day MA | Strong |
| Pattern at converging 50/200 MA | Very strong |
| Pattern far from any MA | Weaker; no structural anchor |
Morris provides statistical testing results. Patterns ranked by reliability:
Key insight: No pattern exceeds ~75% reliability on its own. The edge comes from combining patterns with trend context, support/resistance, and volume.
1. IDENTIFY the prevailing trend (up, down, or sideways).
2. WAIT for price to reach a significant level (S/R, MA, trendline).
3. RECOGNIZE a candlestick reversal or continuation pattern.
4. CONFIRM with volume and/or next-day price action.
5. ENTER on confirmation:
- Bullish: buy above the high of the pattern's final candle.
- Bearish: sell below the low of the pattern's final candle.
6. PLACE stop-loss beyond the pattern's extreme.
7. SET target at next significant S/R level or use risk-multiple (2:1 minimum).
Aggressive entry: Enter at the close of the pattern's final candle (before next-day confirmation). Use only for Tier 1 patterns at major levels with heavy volume.
Conservative entry: Wait for next-day confirmation (price trades above the pattern high for bullish, below the pattern low for bearish). Accepts slightly worse price in exchange for higher probability.
Ultra-conservative entry: Wait for a close above/below the pattern's extreme. Misses some trades but avoids most false signals.
Tier 1 pattern + volume + key level -> Full position size (e.g., 1.0R)
Tier 2 pattern + some confirmation -> Reduced position (e.g., 0.5-0.75R)
Tier 3-4 pattern -> Minimum position or skip (e.g., 0.25R)
| Pattern | Stop Placement |
|---|---|
| Hammer | Below the hammer's low |
| Shooting Star | Above the shooting star's high |
| Engulfing | Beyond the engulfing candle's extreme |
| Morning/Evening Star | Beyond the star candle's extreme |
| Piercing Line | Below the low of Candle 1 |
| Dark Cloud Cover | Above the high of Candle 1 |
Always add a small buffer beyond the pattern extreme to avoid stop-hunting:
An alternative approach: place the stop at pattern_extreme - 1.5 * ATR(14). This
adapts to current volatility and avoids arbitrary fixed-distance stops.
The single most frequent error. A hammer is not a buy signal. A hammer, at support, in a downtrend that has become overextended, with rising volume, followed by a confirmation day β that is a buy signal.
Beginners see a "bearish engulfing" in a screaming uptrend and go short. The pattern is not bearish if the trend is up β it may simply be a pullback within the larger trend. Always identify the higher timeframe trend first.
Setting body_size thresholds to four decimal places based on backtesting is curve-fitting. Candlestick patterns are visual and approximate by nature. Use reasonable thresholds and accept that some subjectivity is inherent.
A bullish engulfing on volume 30% below average is not the same as one on volume 200% above average. Volume validates the conviction behind the pattern.
Candlestick patterns tell you when to enter, not how far price will go. Use separate techniques (measured moves, Fibonacci extensions, prior S/R) for target setting.
Candlestick patterns were designed for daily charts. They can work on weekly or 4-hour charts, but on 5-minute or 1-minute charts, the noise-to-signal ratio increases dramatically. Extend confirmation requirements on lower timeframes.
CONTEXT:
- Stock XYZ has declined from $85 to $62 over 8 weeks.
- The 200-day MA is at $63. The prior pivot low from 6 months ago is at $61.50.
- RSI(14) is at 28 (oversold).
DAY 1 (Monday):
- Large bearish candle: Open $65.00, High $65.50, Low $62.10, Close $62.50
- Volume: 1.2x 20-day average
- Continues the downtrend into the support zone.
DAY 2 (Tuesday):
- Small doji: Open $61.80, High $62.20, Low $61.40, Close $61.90
- Volume: 0.6x average (low volume = indecision, not conviction selling)
- Gaps below Day 1's close. Body is tiny. This is the "star."
- Low touches $61.40, just below the $61.50 pivot support.
DAY 3 (Wednesday):
- Large bullish candle: Open $62.30, High $64.80, Low $62.10, Close $64.50
- Volume: 2.1x average (surge in buying interest)
- Closes well into Day 1's body (Day 1 midpoint = $63.75; close of $64.50 > $63.75).
PATTERN IDENTIFIED: Morning Doji Star at support (Tier 1).
CONFIRMATION CHECKLIST:
[x] Downtrend present (8-week decline, lower highs/lows)
[x] Pattern at significant support (200-day MA + prior pivot)
[x] Oscillator oversold (RSI 28)
[x] Volume surge on Day 3
[x] Day 3 closes above Day 1 midpoint
ENTRY: Buy at $64.90 (above Day 3 high of $64.80, triggered Day 4 morning).
STOP: $61.25 (below Day 2 low of $61.40 minus buffer).
RISK: $64.90 - $61.25 = $3.65 per share.
TARGET: $72.00 (prior consolidation zone), reward = $7.10, R:R = 1.95:1.
POSITION SIZE: With $100,000 account risking 1% ($1,000):
Shares = $1,000 / $3.65 = 273 shares.
Position value = 273 * $64.90 = $17,718 (17.7% of account).
MANAGEMENT:
- Day 5: Price reaches $66.50. Move stop to breakeven ($64.90).
- Day 9: Price reaches $69.00. Trail stop to $67.00 (below Day 8 low).
- Day 12: Price hits $72.00 target. Sell 50%. Trail remaining with 2-ATR stop.
- Day 15: Trailing stop triggered at $70.50. Exit remaining shares.
RESULT: Average exit ~$71.25. Profit = ($71.25 - $64.90) * 273 = $1,733 (1.73R).
A candlestick pattern is a warning, not a command. It tells you something is changing; it does not guarantee the change will follow through.
The trend determines whether you look for bullish or bearish patterns. In an uptrend, look for bullish continuation patterns; bearish reversal patterns at the top only after extended moves. Never fight the trend on a single candle.
Support and resistance give patterns their address. A pattern floating in empty price space has far less significance than one at a well-established level.
Volume is the lie detector. Price can be manipulated in the short term; sustained volume cannot. A pattern without volume confirmation is suspect.
Confirmation transforms a pattern into a signal. The next day's action either validates or invalidates the pattern. Be patient enough to wait for it.
Risk management is not optional. Every candlestick trade must have a predefined stop-loss based on the pattern's structure. The pattern defines the risk; you decide whether that risk is acceptable for your account.
Combine Eastern and Western techniques. Candlestick patterns identify when; moving averages and trendlines identify where; oscillators identify whether conditions are ripe. Together, they form a complete framework.
Simplicity over complexity. Master 8-10 core patterns rather than memorizing 60. The most reliable patterns are the most visually obvious ones: engulfing, morning/ evening star, hammer, and three white soldiers/black crows.
Timeframe matters. Daily charts produce the most reliable signals. Weekly charts produce fewer but more powerful signals. Intraday charts produce more noise and require tighter confirmation rules.
Statistical humility. Even the best patterns under ideal conditions have a 65-75% win rate at most. This means 25-35% of signals will fail. Only consistent risk management turns a probabilistic edge into long-term profitability.
End of implementation specification.