Direct answer
“Engulfing” refers to a price-action setup where the real body of one candle overlaps and “engulfs” the real body of the prior candle. Advanced considerations are mostly about (1) defining the measurement precisely, (2) separating stable mechanics of the candle-body rule from variable conditions like volatility and execution costs, and (3) explicitly managing edge cases where different implementations produce different results.
This article explains how to describe engulfing in a way you can independently verify: what the pattern is, what assumptions any example uses, what can fail, and what you would need to check in your own interpretation.
Mechanism and clear definition
1) Choose the exact rule for “engulf”
A practical engulfing definition needs a clear comparison that you can apply mechanically to OHLC bars (open, high, low, close).
Common building blocks are:
- Candle body: the range from open to close. The direction depends on whether close is above open (bullish candle) or below open (bearish candle).
- Engulfing condition: the body of the second candle overlaps the body of the first candle in the sense that the second body’s boundary extends past the first body’s opposite boundary.
To keep it verifiable, you should specify statements like:
- Bullish engulfing: the second candle is bullish and its body lower boundary is at or below the first candle’s body lower boundary, and its body upper boundary is at or above the first candle’s body upper boundary.
- Bearish engulfing: the second candle is bearish and its body upper boundary is at or above the first candle’s body upper boundary, and its body lower boundary is at or below the first candle’s body lower boundary.
This is a stable mechanical rule, but it is still a definition. Different sources implement details differently, and those differences can change which instances are counted.
2) Decide how to treat equality and “at least”
Edge cases are where ambiguity hides:
- What if the engulfing candle’s body boundary matches exactly (e.g., equal upper boundaries)?
- Does “engulf” require strict inequality, or is equality allowed (“at or beyond”)?
If you allow equality, you may count more patterns; if you require strict engulfing, you may count fewer. Either can be consistent—what matters is that your rule is explicitly written so others can reproduce the same classification.
3) Distinguish bodies from wicks
Many misunderstandings come from mixing wick behavior (high/low extremes) with body overlap. Engulfing definitions are body-based. You should specify whether you ignore wicks except for determining candle direction.
Example of an explicit approach (assumption stated):
- Assume you only use open and close to check engulfing.
- Assume you do not require anything about wicks (for instance, you do not require the engulfing candle to close near its high).
With these assumptions, two charts that have identical open/close series will classify the same engulfing events even if wick shapes differ.
4) Separate “directional intent” from the measurable pattern
It helps to treat engulfing as two parts:
- Pattern classification (measurable from OHLC bodies).
- Interpretation (what you think the pattern “means” in context).
Advanced considerations focus on keeping interpretation from contaminating classification. In other words: first verify the rule identifies the same candles; then you can evaluate, with appropriate caution, whether any additional context improves or harms reliability.
Evidence and example scenarios (with stated assumptions)
Because there is no real-time data here and outcomes vary, the safest way to discuss “evidence” is through example implementations and what you would check.
1) A minimal coding model you can verify
Assume you have time-ordered candles at a fixed timeframe (e.g., one-minute bars, one-hour bars). Define:
- Body boundaries: body_high = max(open, close), body_low = min(open, close).
- Candle direction: bullish if close > open, bearish if close < open.
Then implement two boolean functions:
- bullish_engulfing(t): candle t is bullish AND body_low(t) <= body_low(t-1) AND body_high(t) >= body_high(t-1).
- bearish_engulfing(t): candle t is bearish AND body_high(t) >= body_high(t-1) AND body_low(t) <= body_low(t-1).
This model is intentionally strict about what is measurable. The “advanced consideration” is that many real implementations differ in one line (strict vs inclusive boundaries, handling doji candles, or defining body_high/body_low differently). If you document your choices, others can reproduce your results.
2) How timeframe changes classification
Assumption: you analyze a single timeframe with fixed bar construction. If you change timeframe, engulfing events can appear or disappear because open/close points change when bars are aggregated. This is not a failure of engulfing; it is a reminder that the pattern is tied to the bar definition.
A practical check (verification task, not advice):
- Pick two timeframes and confirm that the set of engulfing instances changes as expected.
- Ensure your classification code uses the same OHLC source and the same timezone handling.
3) Partial overlaps and “near engulfing”
If you interpret “engulfing” loosely (for example, “mostly overlaps” or “wick-to-wick engulfing”), you are no longer using a single reproducible rule.
A safer approach is to avoid fuzzy thresholds unless you define them precisely. For example, you could define a measurable overlap ratio:
- overlap = length of intersection of body intervals.
- require overlap >= X%.
Advanced consideration: choosing X introduces a new parameter and can overfit if you tune it to historical data. If you use such thresholds, keep them transparent so independent verification is possible.
Limitations and risks (what can fail)
1) Costs and execution effects
Even if engulfing is classified correctly, realized outcomes depend on market microstructure:
- spreads (difference between bid/ask),
- slippage (difference between expected and actual fill),
- and any operational constraints that affect when you can enter/exit.
Because these factors vary by provider, instrument, and time, historical “pattern success” does not automatically translate into real results.
2) Data quality and bar construction
Engulfing classification depends on open and close values. Data differences can cause different labels:
- corporate actions or contract rollovers (if relevant for your instrument set),
- timezone alignment and session boundaries,
- missing bars or different feed sources.
Advanced check: verify your OHLC series against another independent dataset for a small sample and confirm that your engulfing labels match under the same rule.
3) Non-stationarity and changing market regimes
Patterns can perform differently across regimes (for instance, high volatility vs low volatility). This is not necessarily a “pattern failure”; it is a reminder that the mapping from pattern occurrence to subsequent returns is not stable.