Skip to main content

Candlestick Patterns

Updated Jun 23, 2026 at 8:47pm

  • 11797af8e6bb Single-Candle Patterns (Doji, Hammer, Marubozu) 1 1,377
  • 11785d35183a Two-Candle Patterns (Engulfing, Harami) 1 1,318
  • 1177a24e6d60 Three-Candle Patterns (Morning/Evening Star) 1 1,056
Tree Key
Expandable — has sub-topics
475Local Id for node
a1b2c3d4Click to see full UUID
3Sub-topics
4Documents
4.7k wordsResearch depth
5Open node
Research Draft Medium 950 words

Candlestick patterns are short formations of one to three price bars — each bar showing the open, high, low and close — read as a snapshot of who held control between buyers and sellers over that interval. The technique originated with 18th-century Japanese rice traders (the lineage usually traced to Munehisa Homma) and was introduced to Western markets by Steve Nison in the early 1990s. A single candle's body (open-to-close) and shadows (the highs and lows beyond it) encode the session's tug-of-war: a long lower shadow with a small body near the top, for example, says sellers pushed price down but buyers reclaimed it before the close. The core tension is that these are momentary sentiment signals, not directional forecasts — their meaning is entirely trend-context-dependent. The same hammer that is a meaningful potential reversal at the bottom of a clean downtrend is noise in the middle of a sideways range. This node defines the family and routes to children grouped by bar count; the specific patterns live in the child nodes.

How they're grouped

Candlestick patterns are conventionally organized by the number of bars that form them, which is how this branch routes:

  • Single-candle patterns — one bar carries the whole signal (Doji, Hammer, Marubozu). The body/shadow geometry alone implies indecision, rejection of a level, or one-sided conviction. → see Single-Candle Patterns.
  • Two-candle patterns — the relationship between two consecutive bars matters (Engulfing, Harami). The second candle's body either swallows or is contained within the first, signalling a momentum flip or a stall. → see Two-Candle Patterns.
  • Three-candle patterns — a small sequence tells a story of exhaustion-then-reversal (Morning Star, Evening Star). → see Three-Candle Patterns.

A second, cross-cutting taxonomy splits patterns into reversal vs continuation signals; most of the named patterns above are reversal patterns, which is where retail attention concentrates.

How they're used in practice

The disciplined use of candlesticks is as a timing-and-context overlay on top of a thesis that already exists, never as a standalone trigger. Three rules recur across the literature:

1. Context first. Nison's most repeated lesson is that no pattern should be read in isolation — a hammer means nothing mid-range; it only carries information when it appears at a tested support/resistance level, a trendline, or after an extended move. The pattern locates where sentiment may be turning; the surrounding structure decides whether to care. 2. Confirmation. Practitioners (and Bulkowski's testing) treat a pattern as actionable only after a confirming bar — typically the next candle closing in the signalled direction. The raw pattern proposes; confirmation disposes. 3. Combination with Western tools. Candlesticks are routinely overlaid on support/resistance, moving averages, volume and momentum indicators (RSI, MACD). The candle gives the entry nuance; the other tools supply the edge.

The swing-trading-SPECIFIC mechanics — exact entry, stop placement below the pattern low, and target — live in the Swing Trading branch, not here.

Standing & evidence (read this honestly)

This is the part a trader most needs and most often skips. The measured standalone edge of candlestick patterns is weak, and the doc should not pretend otherwise:

  • The key academic result, Marshall, Young & Rose (2006, Journal of Banking & Finance), tested candlestick strategies on Dow Jones Industrial Average component stocks (1992–2002) using a bootstrap that generates random OHLC series. They found no statistically significant value — candlestick signals did not beat random trading in liquid US large-caps after accounting for the data's structure. Mixed and sometimes positive results appear in other markets (e.g. some Japanese- and Thai-market studies), so the evidence is not uniformly negative, but the cleanest large-cap-US test is null.
  • Bulkowski's measured base rates are modest and confirmation-dependent. He reports reversal patterns as a group only modestly favouring reversal, with widely-cited individual hit rates (e.g. evening star around the low-70s percent with confirmation, doji closer to a coin-flip on its own). These figures are sensitive to how "success" is defined and to the confirmation rule used — they are not a portable edge.

The honest framing: candlestick patterns are a sentiment/timing-nuance overlay, not a standalone signal. Treat any pattern-only "win rate" with suspicion unless its source, market, period and confirmation rule are stated.

Strengths & limitations

Strengths: fast visual read of intrabar buyer/seller dynamics; useful for refining entry timing once a thesis exists; language-agnostic and applies across timeframes. Limitations: weak measured standalone edge (above); high false-signal rate without confirmation; strongly regime- and location-dependent. The single most common misuse is trading a pattern in isolation — spotting a "hammer" anywhere on a chart and acting on it without checking trend context or confirmation. That is precisely the use the academic evidence finds worthless.

System relevance

Within Delvantic, candlestick reads are an input nuance, not a decision. The Augustus trade-setup agent should never treat a pattern as a standalone entry trigger — it should weight it only when context (trend, level, regime) and confirmation are present, and defer the question of whether the signal actually pays to Cairn's measured track record. Hard caveat for downstream consumers: do not let candlestick "win rates" borrow credibility they have not earned in liquid US equities.

Sources