The full catalogue of classic patterns — Nison's candlestick formations, Bulkowski's chart patterns, harmonics, Elliott and Wyckoff — with how each one forms, what it says about the market, and how it is usually traded. Patterns marked auto-detected are found for you automatically in the screener and on every coin and stock page.
A small body at the top of the candle with a lower wick 2–3× the body, appearing after a decline. Sellers pushed price far down intraday, but buyers rejected the whole move back to the open — demand is stepping in at these levels.
How to trade: Buy on a break of the hammer's high (or the next candle's confirmation); stop below the hammer's low. Only meaningful after a downtrend or at support.
Historical stats (in-sample backtest): 417 trades · 31.4% win rate · avg -0.08R per trade
The identical shape to the hammer — small body, long lower wick — but printed after an advance. The same intraday selling now shows supply appearing inside an uptrend: the first crack in the ceiling.
How to trade: Sell/short on a break of its low with a stop above its high. Context is everything: same form as the hammer, opposite meaning.
Historical stats (in-sample backtest): 445 trades · 31.7% win rate · avg -0.07R per trade
After a decline, a small body sits at the bottom of a long upper wick. Buyers probed sharply higher during the session; even though the close fell back, the probe shows demand testing the ceiling.
How to trade: Wait for a green confirmation candle, then buy its break; stop below the inverted hammer's low.
Historical stats (in-sample backtest): 375 trades · 36.5% win rate · avg +0.05R per trade
The mirror of the inverted hammer, after an advance: a long upper wick shows buyers drove price up and were completely rejected — supply overwhelmed the breakout attempt.
How to trade: Sell on a break of its low; stop above the wick's high. Strongest at resistance or after an extended run.
Historical stats (in-sample backtest): 406 trades · 27.8% win rate · avg -0.20R per trade
Open ≈ close at the very top of the range with a long lower tail. The entire session's selling was absorbed and reversed — a spring-loaded rejection of lower prices, meaningful after a downtrend or at support.
How to trade: Buy on a break of the doji's high; stop below its low.
Historical stats (in-sample backtest): 166 trades · 42.8% win rate · avg +0.21R per trade
Open ≈ close at the very bottom with a long upper tail. Buyers controlled most of the session and lost all of it by the close — a strong rejection at the top, meaningful after an uptrend or at resistance.
How to trade: Sell on a break of the doji's low; stop above its high.
Historical stats (in-sample backtest): 168 trades · 34.5% win rate · avg -0.03R per trade
Open and close nearly equal — a session that ends in a draw. On its own it only signals indecision; after a long trend it warns the driving side is losing conviction.
How to trade: Not a standalone entry. Treat it as a caution flag and trade the confirmation candle that follows.
A small body centred between roughly equal wicks: both sides fought and neither won. A cluster of spinning tops marks a market that is 'thinking' — energy building for the next move.
How to trade: No standalone signal. Use it to tighten stops on existing positions or wait for the range to break.
A candle with no (or almost no) wicks — the open is one extreme and the close the other. One side held total control from bell to bell; typically used as a confirming candle for other setups.
How to trade: Trade in the marubozu's direction, ideally as breakout confirmation; stop beyond its opposite end.
Historical stats (in-sample backtest): 429 trades · 35.0% win rate · avg +0.02R per trade
After a directional drift, a solid candle opens exactly at its extreme (no wick on the entry side) and drives the other way the whole session. The open itself became the floor (bullish) or ceiling (bearish).
How to trade: Enter on the close or the next bar's confirmation; stop just beyond the belt-hold's open.
Historical stats (in-sample backtest): 745 trades · 33.0% win rate · avg -0.03R per trade
A small candle followed by an opposite-colour candle whose body completely swallows it. Control changed hands inside a single session — one of the most reliable two-candle reversals.
How to trade: Enter on the close of the engulfing candle or the break of its extreme; stop beyond the opposite end of the engulfing body.
Historical stats (in-sample backtest): 767 trades · 32.6% win rate · avg -0.04R per trade
A large candle followed by a small opposite-colour candle contained entirely within its body. Momentum abruptly contracted — a pause that often precedes reversal, weaker than an engulfing.
How to trade: Wait for the next candle to confirm direction (see Three Inside). Stop beyond the large candle's extreme.
Historical stats (in-sample backtest): 1068 trades · 34.1% win rate · avg -0.00R per trade
A red candle, then a green one that opens below its low but closes above the midpoint of its body. Sellers gapped the market down and were overrun — deep enough to matter, unlike the weaker thrusting line.
How to trade: Buy on the close or next-bar break; stop below the second candle's low.
The bearish mirror: a green candle, then a red one opening above its high but closing below the midpoint of its body — enthusiasm sold into hard.
How to trade: Sell on the close; stop above the second candle's high.
Two adjacent candles with practically identical highs (top) or lows (bottom) — the same level rejected twice in immediate succession.
How to trade: Trade the rejection: entry on the break of the second candle's opposite side, stop just beyond the shared extreme.
Historical stats (in-sample backtest): 127 trades · 30.7% win rate · avg -0.09R per trade
A candle in one direction, then a gap open beyond the prior open and a candle the other way, bodies never overlapping. Usually news-driven: the single most abrupt sentiment flip in candlestick analysis.
How to trade: Enter in the kicker's direction quickly — there is rarely a pullback; stop inside the gap (its fill invalidates the signal).
A true gap between consecutive candles in trend direction — Japanese analysis treats the gap zone itself as future support (rising) or resistance (falling). Rare on 24/7 crypto majors, common on stocks.
How to trade: Trade continuation while the window stays open; a full gap fill cancels the signal. Pullbacks to the gap edge are the classic entry.
After a strong move, the next candle opens far beyond the prior close and then returns exactly to it — a violent excursion fully neutralised. A weaker cousin of the kicker that warns the trend is tiring.
How to trade: Not a standalone entry; treat as a warning. Aggressive traders fade the trend on the next confirmation candle, stop beyond the counterattack extreme.
A family of weak pullbacks inside a downtrend: the green candle opens below the prior low and closes only at that low (on-neck), just inside the body (in-neck) or below its midpoint (thrusting). Each shows buyers too weak to turn the tide.
How to trade: Bearish continuation setups: short the break of the pullback candle's low; stop above the prior red candle's midpoint.
Three candles: a large red, a small-bodied pause (star, often gapped down), then a large green closing deep into the first body. Panic, indecision, reversal — one of the most reliable bottoming sequences.
How to trade: Buy on the third candle's close; stop below the star's low. The doji-star variant is even stronger.
The mirror top: large green, gapped small-bodied star, then a large red closing deep into the first body — euphoria, hesitation, distribution.
How to trade: Sell on the third candle's close; stop above the star's high.
Three consecutive green candles, each closing near its high and above the last, with small wicks — steady, confident buying. Caution: printed after a long rally it can mark overextension instead of a fresh start.
How to trade: Enter on the third close or a shallow pullback; stop below the first soldier's low.
Historical stats (in-sample backtest): 68 trades · 27.9% win rate · avg -0.19R per trade
Three consecutive red candles stepping lower with small wicks — methodical distribution, the bearish mirror of the soldiers.
How to trade: Short the third close or a weak bounce; stop above the first crow's high.
Historical stats (in-sample backtest): 107 trades · 34.6% win rate · avg +0.01R per trade
A big green candle, up to three small red candles resting inside its range (profit-taking, not reversal), then a new green candle closing above the first one's high — the trend simply caught its breath.
How to trade: Buy the close/breakout of the final green candle; stop below the consolidation's low.
The bearish mirror: a big red candle, small green corrective candles held inside its range, then a new red candle closing below its low.
How to trade: Short the break of the final red candle's low; stop above the correction's high.
A harami plus a third candle closing beyond the first candle's body — the confirmation that fixes the harami's main weakness.
How to trade: Enter on the third candle's close; stop beyond the first candle's extreme.
Historical stats (in-sample backtest): 157 trades · 35.0% win rate · avg +0.01R per trade
An engulfing pattern plus a third candle extending the move — confirmation that the new side kept control.
How to trade: Enter on the third candle's close; stop beyond the engulfing candle's extreme.
Historical stats (in-sample backtest): 1046 trades · 32.4% win rate · avg -0.05R per trade
Red–green–red where the two red candles close at practically the same price: the market bent twice at the same level and refused to break it.
How to trade: Buy the break of the middle candle's high; stop below the shared closing level.
Historical stats (in-sample backtest): 454 trades · 41.9% win rate · avg +0.23R per trade
Three green candles in a row, but each body smaller and upper wicks longer — the uptrend is being sold into even as it rises. An internal-exhaustion warning rather than an outright reversal.
How to trade: Not an entry — tighten stops on longs and watch for a bearish trigger (engulfing, evening star) to act.
Historical stats (in-sample backtest): 242 trades · 33.1% win rate · avg -0.02R per trade
A doji gapped away from both the candle before and after it — the market abandoned an entire price zone in two sessions. Very rare and very strong.
How to trade: Enter in the direction of the second gap on its confirmation; stop beyond the doji.
A gap down in a downtrend, then a green candle that pushes into the gap without closing it — the bounce failed inside the gap zone.
How to trade: Bearish continuation: short when price rolls over inside the gap; stop above the gap's top.
A cluster of small indecisive candles (the tired phase) resolved by one powerful candle breaking the cluster's boundary. Not a canonical Nison pattern, but the everyday anatomy of a consolidation breakout — flags and pennants are its structured cousins.
How to trade: Enter on the impulse candle's break of the consolidation boundary; stop on the other side of the cluster.
Historical stats (in-sample backtest): 31 trades · 41.9% win rate · avg +0.26R per trade
A sharp impulse (the 'pole') followed by a brief, narrow channel drifting against the trend. The market pauses to digest the move; too orderly and shallow to be distribution.
How to trade: Enter on the channel breakout in the pole's direction; stop beyond the far channel edge; target = pole length projected from the breakout.
Historical stats (in-sample backtest): 155 trades · 29.7% win rate · avg +0.15R per trade
Like the flag but the pause converges into a tiny triangle instead of a channel — usually shorter-lived and tighter.
How to trade: Same plan as the flag: breakout entry, stop beyond the opposite boundary, pole-length target.
Historical stats (in-sample backtest): 53 trades · 22.6% win rate · avg -0.44R per trade
A flat resistance line with rising lows pressing into it. Buyers pay up ever higher while sellers defend one level — pressure that usually resolves upward.
How to trade: Buy the horizontal breakout; stop under the last higher low; target = triangle height from the breakout.
Historical stats (in-sample backtest): 68 trades · 39.7% win rate · avg -0.10R per trade
The mirror: flat support with lower highs pressing down into it — supply keeps arriving earlier each bounce.
How to trade: Short the support break; stop above the last lower high; target = triangle height projected down.
Historical stats (in-sample backtest): 92 trades · 32.6% win rate · avg -0.20R per trade
Higher lows and lower highs converging — volatility compresses and the market coils. Direction is genuinely unknown until the break.
How to trade: Trade only the confirmed breakout, ideally on rising volume; stop beyond the opposite trendline.
Price oscillating between two flat, parallel levels — balance between supply and demand until one side is exhausted.
How to trade: Either fade the edges inside the range, or trade the breakout in the direction of the prevailing trend; stop beyond the far boundary.
A trend moving between two parallel sloped lines — an orderly trend with rhythm. The trend side of the channel is the higher-probability trade.
How to trade: Buy the lower line in an ascending channel (sell the upper in a descending one); stop beyond the channel; a channel break often starts a new trend phase.
Historical stats (in-sample backtest): 401 trades · 56.6% win rate · avg -0.01R per trade
A near-vertical run (roughly doubling) followed by a shallow pause giving back less than a quarter of it — Bulkowski's best-performing continuation pattern, typical of speculative low-caps and altcoins.
How to trade: Buy the consolidation breakout; stop under the pause's low; target = full pole projected from the breakout.
A series of J-shaped rounded consolidations repeating up (or down) the trend — rare but well documented by Bulkowski. Not auto-detected: its free-form curve resists a robust geometric definition (our rounding detector covers the closest single-arc case).
How to trade: Enter as each J-curve breaks its rim in trend direction; stop under the curve's base.
Historical stats (in-sample backtest): 123 trades · 39.8% win rate · avg -0.04R per trade
Three peaks — a higher head between two shoulders — with a neckline under the troughs. Each rally attempt is weaker: the textbook picture of demand exhaustion, and statistically among the most reliable reversals.
How to trade: Short the neckline break; stop above the right shoulder; target = head-to-neckline distance projected down from the break.
Historical stats (in-sample backtest): 71 trades · 50.7% win rate · avg +0.02R per trade
The mirror bottom: three troughs with the deepest in the middle, then a break of the neckline above — selling pressure exhausted in stages.
How to trade: Buy the neckline break; stop below the right shoulder; target = pattern height projected up.
Historical stats (in-sample backtest): 40 trades · 27.5% win rate · avg -0.31R per trade
Two roughly equal peaks with a trough between (an 'M'). The second failure to break the high confirms fresh demand is gone.
How to trade: Short the break of the middle trough (the neckline); stop above the second peak; target = pattern height from the break.
Historical stats (in-sample backtest): 46 trades · 45.7% win rate · avg -0.09R per trade
The 'W' mirror: two roughly equal lows — the second test holds on weaker selling, showing supply exhaustion.
How to trade: Buy the break of the middle peak; stop below the second low; target = pattern height projected up.
Historical stats (in-sample backtest): 38 trades · 44.7% win rate · avg +0.30R per trade
Three failed tests of the same ceiling — more information than a double top, hence a stronger signal when the neckline finally goes.
How to trade: Same plan as the double top: short the neckline break, stop above the peaks, height-projected target.
Historical stats (in-sample backtest): 170 trades · 44.7% win rate · avg -0.00R per trade
Three holds of the same floor — each test on weaker pressure. A patient accumulation picture.
How to trade: Buy the neckline break; stop below the lows; height-projected target.
Historical stats (in-sample backtest): 180 trades · 42.2% win rate · avg +0.19R per trade
Both boundaries rise but converge — price grinds higher on shrinking range and typically fading volume. The advance is running out of fuel even while making highs; usually resolves down, even inside an uptrend.
How to trade: Short the break of the lower boundary; stop above the wedge; target = the wedge's height at its origin.
Historical stats (in-sample backtest): 108 trades · 50.0% win rate · avg +0.03R per trade
The mirror: both lines fall and converge — selling pressure exhausting itself. Usually resolves upward.
How to trade: Buy the break of the upper boundary; stop below the wedge; target = wedge height.
Historical stats (in-sample backtest): 164 trades · 53.0% win rate · avg +0.02R per trade
A smooth U-shaped base back to a prior high, then a small drifting pause (the handle) on the right rim — gradual accumulation with a final shakeout of weak hands before the breakout.
How to trade: Buy the break of the handle's high; stop under the handle's low; target = cup depth projected from the rim.
Historical stats (in-sample backtest): 80 trades · 51.2% win rate · avg +0.05R per trade
A rounded dome then a small upward-drifting handle before the breakdown — distribution disguised as resilience.
How to trade: Short the break of the handle's low; stop above the handle; target = dome depth projected down.
Historical stats (in-sample backtest): 66 trades · 36.4% win rate · avg -0.20R per trade
Three phases: an orderly lead-in trend, a speculative acceleration pulling twice as far from the trendline (the bump), then the run — a collapse back through the line. The anatomy of a blow-off.
How to trade: Enter when price breaks back through the lead-in trendline; stop beyond the bump's extreme; first target = the bump's full height beyond the line.
Two adjacent candles with very long wicks pinned to the same extreme, standing clear of everything around them — a violent two-bar double rejection, classically strongest on weekly charts.
How to trade: Enter on the break of the pair's opposite side; stop beyond the shared spike extreme.
Historical stats (in-sample backtest): 74 trades · 44.6% win rate · avg +0.30R per trade
Like the pipe but with one calm bar between the two spikes. Unlike a double top the spikes sit close together and need not match exactly.
How to trade: Same plan as the pipe: break of the opposite side, stop beyond the spikes.
Historical stats (in-sample backtest): 64 trades · 57.8% win rate · avg +0.46R per trade
The anti-triangle: higher highs AND lower lows — expanding swings, rising emotion, no agreement on value. Common near major tops; hard to trade because the boundaries keep moving away.
How to trade: Best treated as a warning to reduce size. Aggressive traders fade the extremes with wide stops beyond the diverging lines.
A megaphone with one flat side — a hybrid of triangle and broadening formation; the flat side is the level the market keeps defending. Not auto-detected separately: our generic broadening detector flags the expanding structure.
How to trade: Trade the break of the flat side; stop inside the formation; expect volatility.
Swings first expand (megaphone) then contract (triangle), tracing a rhombus. Rare, but regarded as a reliable reversal — especially as a top.
How to trade: Enter on the break of the contracting side against the prior trend; stop beyond the diamond's midline; target = the diamond's maximum height.
Historical stats (in-sample backtest): 129 trades · 44.2% win rate · avg -0.01R per trade
A slow, smooth arc with no distinct handle — the balance of supply and demand rotating gradually. Forms over long periods, typically on higher timeframes.
How to trade: Enter as price clears the rim of the saucer; stop inside the arc; target = the arc's depth.
Historical stats (in-sample backtest): 280 trades · 48.6% win rate · avg +0.05R per trade
A gap strands a cluster of candles away from the main move, then an opposite gap abandons them — an island on the chart. Rare (needs gaps, so mostly stocks), usually on a volume spike, and decisive.
How to trade: Enter in the direction of the second gap immediately; stop inside the island. A close back inside the gap voids the signal.
Three successive symmetric pushes to a new extreme with Fibonacci-proportioned corrections between — closer to harmonic/wave analysis than classical charting. Exhaustion by repetition.
How to trade: Fade the third drive once it stalls; stop beyond its extreme; first target = the last correction level.
Historical stats (in-sample backtest): 185 trades · 57.8% win rate · avg +0.13R per trade
An XABCD harmonic where D completes at ~0.786 of XA with matching AB=CD symmetry. Defines a precise reversal zone (PRZ) instead of a vague area.
How to trade: Enter at the D-point zone with a stop just beyond X; targets at the 0.382/0.618 retracements of AD.
A harmonic with a deep D completion (~0.886 of XA) after a shallow B — the PRZ sits close to the origin, offering tight risk.
How to trade: Enter at D; stop beyond X; scale out at the 0.382/0.618 retracements of the CD leg.
The D-point extends BEYOND X (1.27–1.618 of XA) — a terminal overshoot pattern that catches exhaustion outside the prior range.
How to trade: Enter at the D extension zone; stop beyond the 1.618 extreme; targets back inside the XA range.
The most extended harmonic: D completes at ~1.618 of XA after a violent CD leg — designed to catch capitulation extremes.
How to trade: Enter at the 1.618 zone; tight stop beyond it; targets at the 0.382/0.618 of the CD leg.
Not a pattern but a full market model: five waves with the trend, three against, each wave fractally nested. Full nested counts are notoriously subjective, but the three HARD rules are mechanical — getcoinsight detects completed rule-consistent impulses (experimental, low reliability) and leaves degree labeling to you.
How to trade: Most practical uses reduce to: join wave 3 (the strongest), and fade wave 5 divergence. Both require independent confirmation.
A five-point channel where wave 5 overshoots the 1–3 line (the sweep), projecting a precise target along the 1–4 line — the key is channel symmetry, not Fibonacci ratios.
How to trade: Enter at the wave-5 overshoot; stop beyond it; target = the 1–4 line projection (EPA).
Historical stats (in-sample backtest): 230 trades · 51.3% win rate · avg +0.16R per trade
A phase model of how large players build positions inside a range: Preliminary Support, Selling Climax on panic volume, Automatic Rally, Secondary Tests on shrinking volume, a Spring (false breakdown), then Sign of Strength and Last Point of Support before markup. The volume signature on each event is as important as the price shape.
How to trade: The classic entries are the Spring and the LPS pullback after a Sign of Strength; invalidation = losing the range low on expanding volume.
Historical stats (in-sample backtest): 32 trades · 31.3% win rate · avg -0.22R per trade
The mirror scheme for tops: Preliminary Supply, Buying Climax, Automatic Reaction, Secondary Tests, an Upthrust After Distribution (false breakout), then Sign of Weakness and Last Point of Supply before markdown.
How to trade: Short the UTAD failure or the LPSY bounce; invalidation = reclaiming the range high on strong volume.
A false break under the range low that is bought back immediately — the shakeout that transfers coins from late sellers to strong hands. Ideally on a volume spike followed by a quiet successful re-test.
How to trade: Buy the reclaim of the range low (or the quiet re-test); stop below the spring's wick; target = the opposite side of the range.
Historical stats (in-sample backtest): 126 trades · 42.9% win rate · avg -0.04R per trade
A false break above the range high that is sold back inside — the bull trap that fills large sell orders at the top. The bearish twin of the spring.
How to trade: Short the failure back under the range high; stop above the upthrust's wick; target = the range low.
Historical stats (in-sample backtest): 117 trades · 47.0% win rate · avg +0.02R per trade
Educational content, not financial advice. Pattern statistics vary by market, timeframe and regime — classical theory (Edwards & Magee, Bulkowski) requires rising volume on the breakout as confirmation; without that filter the share of false signals grows sharply. Always use protective stops.