Building on 01 · Candlestick Patterns, this article pushes candlesticks from "recognizing shapes" to "reading the language": how to quantify wick ratios, how to confirm and falsify combination patterns, how timeframes resonate, how volume endorses the candles, and how to read them differently across markets.
💡 Master Principle
Remember one master principle first: a single candle is a "word", a combination pattern is a "sentence", volume is the "tone", and the timeframe is the "context". Only when all four are present is a candlestick reading complete.
1. Wick Language
1.1 What a Wick Is
A candle consists of three parts: upper wick, body, and lower wick. The body is "the battlefield outcome from open to close"; the wicks are "the extreme prices that got knocked away and pulled back during the battle":
text
│ ← high
┌──┴──┐ ← upper wick (spike smashed back)
│ body│ ← between open and close
└──┬──┘
│ ← lower wick (probe pulled back)
│
└ ← low
A wick is essentially "a zone price probed but failed to hold". The longer the upper wick, the more selling pressure up above knocked price down; the longer the lower wick, the more buying support down below held price up. Wicks are always "rejected attempts".
1.2 Deep Reading of Long Upper Wicks by Location
| Location | Meaning | Signal strength |
|---|---|---|
| Mid-downtrend | A bounce smashed back; bears still in command; end-of-bounce signal | Weakly bearish |
| End of downtrend (at lows) | First large-scale bull counterattack; even though it didn't hold, selling pressure may be exhausted — confirm reversal with a next-day bullish candle | Leaning bullish (pending confirmation) |
| Mid-uptrend | Normal turnover/shakeout; continues rising after retesting the MAs; often covered by the next day's bullish candle | Neutral |
| End of uptrend (at highs) | Buyer exhaustion and heavy overhead supply; classic topping signal (shooting star) | Strongly bearish |
| Top of a long-term box | A record of repeated failed spikes; reinforces the box-top resistance | Neutral-bearish |
💡 Location Decides Meaning
The rule: the same long upper wick is "the bulls' first word" at the end of a decline and "the bulls' last word" at the end of a rally — location decides meaning; the shape by itself counts for nothing.
1.3 Deep Reading of Long Lower Wicks by Location
| Location | Meaning | Signal strength |
|---|---|---|
| Mid-uptrend | Quick absorption during a pullback; shakeout in nature; continues rising after the retest | Neutral |
| End of uptrend (at highs) | Hanging man: violent intraday selling temporarily absorbed, but the selling pressure is real | Strongly bearish (pending next-day confirmation) |
| Mid-downtrend | Oversold bounce-back; decline resumes afterward; the "falling-knife catching" scene | Weakly bullish (most traps) |
| End of downtrend (at lows) | Bear attack failed with strong buying support; classic bottoming signal (hammer) | Strongly bullish |
| Key support level (near prior low/neckline) | A "stress-test record" of the support; if it holds, the support gains credibility | Leaning bullish |
Beware long lower wicks mid-downtrend: in a downtrend, every long lower wick looks like a bottom, but most are just "dead-cat bounces" within the decline. The criterion is whether the next day can close bullish on volume above the prior day's body.
1.4 Quantifying Wick Ratios
To upgrade wick analysis from "feel" to "measurement", classify a candle by the ratios of upper wick (U) / body (B) / lower wick (L):
text
Ratio rules (body = 1 unit):
upper wick / body < 0.25 → no upper wick (bulls fully control the close)
upper wick / body 1–2× → long upper wick (spike rejected)
upper wick / body ≥ 2× → extreme upper wick (strong rejection)
lower wick: mirror reading
| Upper : body : lower | Name | Reading |
|---|---|---|
| 0 : 1 : 0 | Marubozu (bullish/bearish) | One side completely dominant; direction set by body color |
| 0 : 1 : ≥2 | Hammer / hanging man (same shape; location names it) | Strong support below; bullish at lows, bearish at highs |
| ≥2 : 1 : 0 | Shooting star (highs) / inverted hammer (lows) | Strong overhead supply; bearish at highs, bullish at lows pending confirmation |
| ≥1 : 1 : ≥1 | Long-legged doji | Violent bull-bear disagreement; direction handed to the next day |
| 0 : 1 : 0 with a huge body | One-price bar / full-body candle | Extreme one-sided move, usually untradeable |
Practical mnemonics:
- A wick ratio ≥ 2× the body is worth attention; 1–2× is just "ordinary fluctuation";
- Both wicks long = maximum disagreement; such candles often precede big moves (direction decided by whichever side brings the volume);
- A candle with tiny wicks and a huge body (marubozu) represents one-sided force — it is "trend language", not "reversal language".
2. Combination Patterns, Advanced
A single candle can only express "the bull-bear battle within one bar"; combination patterns express "the handover process of power". Every pattern in this section comes with: formation → confirmation → invalidation — all three steps, none skippable.
2.1 Morning Star / Evening Star
Formation (three candles):
- Morning star (bottom): ① big bearish candle (decline) → ② small body or doji (either color, gaps allowed, representing a bull-bear standoff) → ③ big bullish candle (closing more than halfway into candle ①'s body).
- Evening star (top): exact mirror — ① big bullish candle → ② small body/doji → ③ big bearish candle penetrating candle ①'s body.
text
Morning star (bottom): Evening star (top):
│ ┌──┐
┌─┴─┐ ┌─┐ │ ① │ ← big bullish candle
│ ① │ │ ② │ ← doji/small body └─┬─┘ ┌─┐
└─┬─┘ └─┘ │②│ │ │ ← small body/doji
│ ┌─┐ └─┘ │ ③ │ ← big bearish candle
(decline)│ ③ │ ← big bull candle deep into ①'s body └─┬─┘
└─┘ (often with volume expansion) (rally)
Confirmation:
- Candle ③ must close more than halfway into candle ①'s body (the deeper, the more effective; engulfing candle ① upgrades it to a stronger reversal);
- For a morning star, candle ③ must come on expanded volume; for an evening star, expanded volume on candle ③ raises certainty;
- Candle ② being a doji (the smaller the body) is more reliable than an ordinary small body — the more complete the standoff, the more sudden the reversal.
Invalidation:
- Candle ③ only bounces modestly (less than halfway into candle ①'s body) — the pattern downgrades to a "continuation signal" and the trend most likely resumes;
- After an evening star, price goes sideways for days without falling, then makes a new high — the top pattern was digested sideways;
- It appears at a non-key location (over 70% of morning stars mid-trend are continuations) — mistaking continuation for reversal.
2.2 Three White Soldiers / Three Black Crows
Formation (three candles):
- Three white soldiers: three consecutive bullish candles with progressively longer bodies, each closing higher day over day, each closing in the upper part of the prior body (no long wicks).
- Three black crows: the mirror — three consecutive bearish candles with progressively longer bodies, each closing lower day over day.
text
Three white soldiers (bottom/early trend): Three black crows (top/late trend):
┌──┐ ┌──┐
│① │ │③ │
┌─┴──┴──┐ ┌─┴──┴──┐
│ ② │ │ ② │
┌─┴────────┴──┐ ┌─┴────────┴──┐
│ ③ │ │ ① │
└─────────────┘ └─────────────┘
closes step higher (bullish bodies lengthen) closes step lower (bearish bodies lengthen)
Confirmation:
- The three bodies grow day over day (accelerating momentum), with very short wicks;
- They appear after a clear decline/rally (soldiers at the bottom, crows at the top);
- Three white soldiers with gently rising volume, ideally near key support.
Invalidation:
- A long upper wick on the third candle (soldiers) / long lower wick (crows) means momentum is already overdrawn — often reverses the next day: "soldiers with wicks, turn and shut the door";
- Bodies progressively shrinking "soldiers" (they should lengthen) — decaying momentum, can fail at any time;
- Appearing mid-range: soldiers followed by sideways drift and then a fall is just an ordinary bounce inside the box;
- If a quick volume-dry stabilization follows the crows, it was panic inertia rather than a trend reversal — don't chase the short.
2.3 Tweezer Bottom / Tweezer Top
Formation: Two (or several) consecutive candles with identical lows (tweezer bottom) or identical highs (tweezer top), like the two prongs of tweezers resting on the same level. The two bodies may point in opposite directions; shape and size are unconstrained.
text
Tweezer top: Tweezer bottom:
┌────┐ ┌────┐
│ ① │ │ ② │ ← two prongs on the same level
└─┬──┘ └─┬──┘ (identical highs)
└──┐ ┌──┘ ┌──┐ ┌──┐
└─┘ │ │ │ │ ← two prongs on the same level
└┬─┘ └─┬┘ (identical lows)
└──┐ ┌──┘
└─┘
Confirmation:
- The prong level must be a key level (prior high/low, round-number gate, near a moving average) — the point of tweezers is "two attacks on the same price, both failed";
- The strongest signal comes when the second candle is itself a reversal candle (the second prong carries a long wick, like a shooting star/hammer);
- Add volume: expanding volume on the second attack that still gets repelled = real selling/buying pressure; shrinking volume may just mean nobody showed up.
Invalidation:
- The two prongs are only "roughly equal" (differing by more than 0.5%–1%) — not tweezers, just an ordinary pullback;
- A third attack breaks straight through the prong level (two prongs become three) — pattern failed — tweezers are only a reversal when it is "exactly twice";
- Appearing mid-move at a non-key level — just ordinary parallel candles inside a range, with no directional meaning.
2.4 Dark Cloud Cover / Piercing Line
Formation (two candles):
- Dark cloud cover (top): ① big bullish candle → ② a big bearish candle opening higher, closing deep below the midpoint of candle ①'s body (but not below its open).
- Piercing line (bottom): mirror — ① big bearish candle → ② a big bullish candle opening lower, closing deep above the midpoint of candle ①'s body.
text
Dark cloud cover (top): Piercing line (bottom):
┌───────┐ ┌─┐
│ ①bull│ │ │ ← ② opens low, big bull candle
│ │ ┌──────┐ │ │
└───────┘ │ ②bear│ ← opens high, closes low │ ② │
│ │ close deep into ①'s body ┌┴─┴─┐
│ ← crosses the midline │ │ ①bear│
└───────────┘ └─────┘
Confirmation:
- Candle ② must gap open (the clearer the gap, the stronger) and close past candle ①'s body midline — stopping at the 1/3 mark is a "failed pierce" and halves the signal;
- Dark cloud cover / piercing line must appear after a distinct rally/decline;
- Candle ② comes with volume (dark cloud on volume = real distribution at the top; piercing on volume = real absorption at the bottom).
Invalidation:
- The close reaches less than 1/3 into candle ①'s body = "no pierce" — just an ordinary pullback/bounce; the pattern does not stand;
- The day after a dark cloud cover closes bullish at a new high (outside reversal back), the pattern fails — candle ② was just a shakeout;
- Appearing mid-box with the next day still oscillating inside — no directional meaning;
- Difference from engulfing: engulfing requires fully wrapping candle ①'s body; dark cloud/piercing only requires crossing the midline — a lower bar to clear, but also a weaker signal.
3. Multi-Timeframe Candle Confluence
💡 Confluence and Conflict
Candlestick patterns can appear on multiple timeframes at once. Confluence = the big timeframe sets direction, the small one times it; conflict = defer to the big timeframe and downgrade the small-timeframe signal.
3.1 Weekly vs. Daily Combination Patterns
| Weekly (big TF) | Daily (small TF) | How to handle |
|---|---|---|
| Uptrend (no topping pattern) | Evening star appears | Just a daily pullback; buy dips at support; no panic |
| Weekly topping pattern (evening star/engulfing) | Morning star appears | The daily bounce is just a pullback off the weekly top; the bounce is a chance to reduce |
| Weekly bottom pattern (hammer/engulfing) | Three white soldiers | Best confluence: big-TF turn + small-TF momentum; a prime participation setup |
| Weekly bottom pattern | Three black crows | The final decline at the daily level; wait for daily stabilization — don't front-run |
| Weekly sideways | Any daily pattern | Pattern downgrades to "oscillation inside a box"; signal credibility halved |
💡 Core Rule
Core rule: the big-timeframe pattern decides "whether it is worth participating"; the small-timeframe pattern decides "when to participate". A small-timeframe signal can never overturn the big-timeframe structure — it can only pick the entry point within that structure.
3.2 Pattern Reconstruction Across Timeframes
The same stretch of price action "grows into" different patterns on different timeframes:
text
Daily view: Weekly view:
╱╲ ╱╲
╱ ╲ ╱ ╲ ╱╲
╱ ╲╱ ╲ ╱ ╲
╱ ╲ ╱ ╲
╱ ╲ ╱ ╲
(looks like a head and shoulders top?) (just a rising wedge / pullback?)
- Patterns are timeframe-dependent: a "head and shoulders top" on the daily may be just a normal pullback within a big weekly uptrend; a weekly hammer may be an entire downleg when viewed on the daily;
- Scaling up the timeframe reduces noise and slows signals: three white soldiers on the 30-minute chart happen almost daily and mean nothing; a hammer on the monthly chart appears once in years and carries enormous weight;
- Reconstruction method: when switching timeframes, first verify whether the "pattern boundary" you saw is still the same line on the bigger timeframe. The same neckline/support sits at different places on different timeframes — a pattern only means something inside the timeframe that defines it.
Common errors:
- Going all-in on a "morning star" seen on the 15-minute chart — small-TF patterns match small-TF moves; at most a quick trade;
- Using a weekly pattern to guide daily trading without waiting for daily confirmation — like aiming a cannon at a distant target and firing with your eyes shut;
- Too many timeframes create "there is always confluence": watch 5 timeframes and one of them will always show the signal you want — keep only 2–3 (e.g., 15m/daily/weekly), and always write down "what the big-TF structure is" before discussing the small TF.
4. Candles and Volume Together
Candles describe "what happened to price"; volume answers "how real it was". A candle pattern without volume endorsement can be treated as nonexistent on illiquid instruments and in thin sessions.
4.1 Pattern + Volume Confirmation Rules
| Pattern type | Correct volume | Dangerous volume |
|---|---|---|
| Bottom reversal (hammer/morning star/piercing) | Reversal candle on expansion (real absorption) | Contraction (merely "can't fall", not "someone is buying") |
| Top reversal (shooting star/evening star/dark cloud) | Reversal candle on expansion (real distribution) | Contraction (supply not exhausted; the fall won't complete) |
| Continuation (three soldiers/three crows) | Gently rising volume | Blow-off volume (momentum spent in one day) |
| Breakout type (tweezer-top break, outside reversal) | Breakout day ≥ 1.5× average volume | Expansion with stalling / spike-and-fade |
| Pause type (hammer during a pullback) | Contracting pullback + expanding bounce | Expanding decline (trend reversal alarm) |
text
The correct posture for a bottom reversal:
price│ ┌─┐
│ │ │ ← expanding bullish candle confirms
│ ┌─┴─┴─┐
│ │hammer│ ← may appear on contraction first (supply exhausted)
└──┴─────┴──────→
volume│ █
│ ███ ← the confirming candle's expansion is "decisive evidence"
└──────────────→
4.2 Three Volume Iron Rules
- For reversals, watch the confirmation day's volume, not the pattern day's: the hammer itself may come on contraction, but the next-day confirming candle must expand — a next-day bullish candle on contraction is just a weak bounce.
- Expansion only counts at a "key level": an expanding bullish candle mid-box is noise; an expanding bullish candle at a prior low/neckline/moving average is a signal — the value of volume = location × magnitude.
- Act only when pattern + volume + location all resonate: shape without volume (weak signal), volume without location (noise), location without shape (ordinary fluctuation) — each is missing a dimension. With all three present, the win rate is worth your money.
5. Candlesticks Across International Markets
Candlesticks were born in the Japanese rice market, but when transplanted into different markets, "market structure" warps what wicks and patterns mean. Changing markets requires changing the frame of reference — the most easily ignored rule of advanced usage.
5.1 US Stocks: The Most "Textbook" Market
- Strengths: superb liquidity, continuous trading hours (including pre/post-market), balanced participant structure — candle patterns track the textbook most closely and wicks carry real battle meaning;
- Traits: the close (especially the closing auction) carries enormous weight; closing patterns (where the close sits within the body) are more reliable than intraday ones;
- Watch out: pre/post-market gaps disrupt candle continuity (open-high-fade patterns are common), and event-driven one-day mega-volume candles (earnings, Fed meetings) have their wick meanings polluted by event noise — do not read them as ordinary technical candles.
5.2 Crypto: 24-Hour Continuous Trading
- Continuous candles: 24 × 7 with no breaks, no "artificial cut points" between candles — the traditional "open" concept is weakened (no psychological anchor at the day's start), and wick meaning depends more on the intraday session (Asian/European/US sessions differ sharply in trade density);
- Wick hunts: in thin-liquidity windows (early morning, weekends), wicks are often "wick hunts" (instant spikes pierced and pulled back); such wicks do not represent real bull-bear battles, only liquidity gaps — hence the false-signal rate of "long lower wick = bottom" in crypto is far higher than in US stocks;
- Gaps: crypto gaps are almost always "hourly-scale", gapping on minute/hour charts and filling quickly; daily gaps are extremely rare; applying A-share gap theory (unfilled for days = valid) to crypto fails completely;
- 7×24 exceptions: exchange maintenance, chain migrations, and major regulatory events can create "artificial gaps" whose fill behavior is unpredictable.
5.3 Forex: No Centralized Exchange
- Continuous but not around-the-clock: forex closes Friday and reopens Sunday — a real weekend discontinuity; Monday's opening candles often carry weekend position adjustments, so their patterns are less reliable than mid-week ones;
- No single matching point: forex is a decentralized interbank quote market with no unified "volume" caliber; volume-based confirmation (Section 4) is nearly unusable directly — rely mainly on price action itself;
- Session-driven: London/New York/Asia sessions have entirely different volatility characters; when one candle spans sessions its wick meaning is not comparable — pattern analysis should proceed by "session slices" rather than "calendar days".
5.4 Holiday Gaps (A-Shares / HK Stocks)
- Long-holiday gaps: gaps around Chinese New Year / National Day in A-shares do get filled in hindsight, but the timing is unpredictable (possibly months later);
- Handling principle: top/bottom patterns that form right before a long holiday deserve a question mark — overnight information accumulated over the holiday is released all at once; once the pattern boundary is breached, defer to the actual post-gap price action rather than stubbornly clinging to the pre-holiday pattern call;
- Christmas / year-end liquidity drought (US/HK stocks): candles in thin liquidity tend to run long wicks and false signals multiply — cut the frequency of pattern trading accordingly.
6. Limitations of Candlesticks
6.1 Success Rates of the Same Pattern Across Markets
| Pattern | Mature equity markets (US/HK blue chips) | A-shares | Crypto / small caps | Forex |
|---|---|---|---|---|
| Hammer bottom | Fairly high (rational participants, thorough turnover) | Medium (emotional + price-limit interference) | Leaning low (many fake wick-hunt wicks) | Medium (session-dependent) |
| Engulfing reversal | High | High (a classic A-share pattern) | Medium (needs volume) | Medium |
| Evening star top | High | Medium (one-price/limit-up days can't form one) | Medium | Leaning low (extreme trend inertia) |
| Doji regime change | Medium | Leaning low (one-price bars swallow dojis) | Medium | Leaning low |
| Gap theory | Modest | High (price limits + overnight system) | Extremely low (almost no daily gaps) | Low (24h continuous) |
💡 Conclusion
Conclusion: there is no candle pattern that "fits all markets". Every pattern embeds assumptions about "market structure" (continuous trading, rational turnover, no price limits); when the structure differs, pattern validity must be recalibrated.
6.2 Risks of Mechanical Application
- Hindsight confirmation bias: looking back at a chart, every top "happens" to have a shooting star and every bottom "happens" to have a hammer — because you only remember the confirmed ones and forget the failures. In real statistics, single-candle patterns mostly win 50%–60% of the time — barely better than a coin flip.
- Subjectivity of pattern recognition: the same candle is a "long upper wick" to one person and "ordinary fluctuation" to another; there is no objective standard for wick thresholds or body sizes, and different statistical calibers produce wildly different conclusions.
- Wrong location is the biggest killer: applying "a shooting star at the end of a rally" to "a bounce mid-decline" inverts the meaning entirely — always ask location first, shape second.
- Candles have no time dimension: after a hammer, the reversal may come in 1 day, 1 week, or 1 month — candles never tell you when it cashes out; you need a stop-loss to absorb the "right signal, wrong timing" risk.
- Even the strongest pattern needs a stop: even a textbook triple resonance (big-TF pattern + expansion + key level) is only a probabilistic edge, not a certainty. Always manage positions by the rules of the 07 · Trading Systems chapter.
⚠️ Risk Warning
Candlesticks are "a language describing past bull-bear battles", not "a crystal ball predicting the future". The same pattern's success rate varies hugely across markets, timeframes, and locations; wicks and patterns generate many false signals in thin-liquidity windows (crypto early mornings, US pre-market, around long holidays); no candlestick signal constitutes a buy/sell instruction — treat patterns as probabilistic edges, always combined with position management and stop-loss discipline, and never abandon risk management or bet the farm with leverage just because "a classic pattern appeared on the chart". This article is for educational purposes only and does not constitute investment advice.