Skip to content

Dragonfly Doji

Patterns candlestick indecision bullish

A Doji variant with a long lower wick, a very small body located near the high of the range, and little or no upper wick. It records a successful defense of lower prices and carries bullish implications after a decline.

Visual Example

Dragonfly Doji: small body near the high with a long lower wick and virtually no upper wick. Annotation highlights the lower support test.

Synthetic ideal matching TA-Lib CDLDRAGONFLYDOJI (long lower shadow + open\approx close\approx high). Generated 2026-05-31 IST via docs/gen_candle_previews.py.

Description

The Dragonfly Doji appears when sellers drive price sharply lower during the session but buyers step in aggressively and push the close back near the open/high. The long lower shadow is the visual footprint of that successful defense.

Most powerful after a downtrend or at support. As with all single-candle patterns, confirmation (next bar close above the Dragonfly high) is required for high-confidence use. It serves as an excellent sparse feature in ML models that already ingest Market Structure bias or Ehlers cycle regime.

Formula / Specification

Recognition Rules (exact implementation in QuantWave / TA-Lib CDLDRAGONFLYDOJI):

  1. Body length negligible (Doji condition).
  2. Lower shadow (min(open, close) − low) long relative to body/range.
  3. Upper shadow (high − max(open, close)) negligible.
  4. Output follows TA-Lib convention.
  5. Single-bar stateless recognition after history buffering.

Parameters

Parameter Default Description
(none) — Pattern recognition only; no tunable parameters.

Usage Examples

(Identical surface examples as Doji/Gravestone, substituting CDLDRAGONFLYDOJI / .ta.cdl_dragonflydoji(...).)

Streaming (Rust / Python / Polars) — see Doji page for pattern; only the type/column name changes. Bit-identical parity holds.

Edge Cases & Limitations

  • First bar yields 0.
  • Can be ignored in strong downtrends without confirmation.
  • Low-range bars distort shadow significance; use ATR filter.
  • Highest edge when occurring at structure support after established bearish bias.
  • Confirmation bar strongly advised.

Boundary Behavior

Condition Behavior
Warm-up Pattern functions emit 0 (no pattern) until enough bars exist.
period > len Short series returns all zeros (no pattern detected).
NaN inputs Bars with NaN OHLC are treated as no pattern (0).
Invalid params N/A for most candlestick patterns.
Empty data Empty input returns an empty integer series.

Sources & References

Primary Source: TA-Lib CDLDRAGONFLYDOJI via quantwave-core/src/indicators/pattern.rs.

Visual: docs/gen_candle_previews.py 2026-05-31 IST.

Context: Nison (1991) psychology only; MQL5 PA for confluence usage.

Provenance: Next<T> + Polars parity contract.