Marubozu candlestick: full, opening and closing variants

Compare full, opening and closing marubozu

TradingView describes the full white marubozu as having no shadow at either end. CandleScanner also distinguishes opening and closing variants, where one shadow can remain and is shorter than the body. The names refer to the open or close end of the candle, not to the opening or closing hour of an exchange. Green and red below mean close above open and close below open, regardless of chart colors.

• For the one-shadow variants, require the remaining shadow to be shorter than the body. A candle with a dominant opposite wick is a different shape.
• A full marubozu also meets both shadowless-end conditions. If labels must be exclusive, classify full first and reserve opening or closing labels for candles with one remaining shadow.

Measure body proportion and shadow tolerance

Use range R = high minus low; body B = the absolute difference between close and open; upper shadow U = high minus the larger of open and close; lower shadow L = the smaller of open and close minus low. Skip zero-range bars before division, and require a nonzero body. B/R measures how much of the range belongs to the body. It does not measure whether that body is unusually large compared with earlier candles.

• Illustrative near-full rule: require B/R of at least 95%, with both U and L no larger than one minimum tick. All three conditions must pass.
• Illustrative opening or closing rule: allow at most one tick at the required shadowless end, require B/R of at least 60%, and require the other shadow to be shorter than B.
• These 95%, 60% and one-tick values are study choices, not universal definitions or TradingView defaults. Keep a zero-tolerance mode when you want the strict pattern.
• A body occupying 95% of a tiny range can still be tiny. Add a separately documented size filter if your study needs a long candle; do not hide that filter inside the shape name.

Why the minimum tick changes near-matches

TradingView exposes the minimum recorded price increment through syminfo.mintick. A one-tick tolerance therefore changes with the symbol. Suppose the tick is 0.01 and a hypothetical bar has open 100.00, high 101.01, low 99.99 and close 101.00. Both shadows are 0.01 and the body is 1.00 out of a 1.02 range, or about 98.04%. It passes the illustrative near-full rule but fails the strict no-shadow definition.

• For a much narrower hypothetical bar, open 100.00, high 100.03, low 99.99 and close 100.02, both shadows are still one tick. Its body occupies only 50% of the 0.04 range, so it fails the 95% body rule.
• Do not replace one tick with a fixed 0.01 price allowance across every instrument. Record the symbol, tick size, timeframe and session with the result.
• Use the OHLC values rather than apparent screen pixels. Chart zoom can make a small shadow disappear visually without changing its measured size.

Read the three shapes on the same scale

These independent bullish examples keep the open and close at 100 and 102. Adding a shadow changes the classification while leaving the two-point body unchanged. The opening and closing examples each have a body-to-range ratio of two-thirds, so they pass the illustrative 60% variant rule but fail the 95% near-full rule.

Separate continuation from a volatility burst

A bullish marubozu candlestick pattern ends above its open, but OHLC does not prove that every intervening trade moved upward. Price can revisit levels inside the body. A wide candle can also reflect a short burst of activity that later reverses. Shape alone cannot distinguish orderly continuation from that outcome.

• Define trend before examining the following candles. One illustrative uptrend rule is a close above a rising 20-bar moving average; reverse the conditions for a downtrend.
• To study unusually wide bars, compare range with a fixed prior volatility measure, such as the preceding bar's 14-period ATR. Choose the cutoff before measuring outcomes.
• Keep shape, trend and volatility filters separate. A filtered result does not describe every marubozu, and this page establishes no win rate or preferred timeframe.

Confirm the bar before confirming continuation

An unfinished candle can gain a shadow or lose its body as price changes. Wait for its close to classify the final shape. TradingView documents barstate.isconfirmed as true on historical bars and on the closing update of a realtime bar. That chart-bar condition does not work inside request.security(); a higher-timeframe detector needs separate confirmation handling.

• For an illustrative continuation study, freeze the completed marubozu high and low, then observe the next three completed bars. A close above the high confirms an upward break; reverse the test for a downward break. If the required directional break does not occur, record expiry.
• A wick above the high followed by a close inside does not meet that close-based rule. A qualifying close followed by a reversal is still a failed continuation outcome, not a reason to erase the signal.
• The three-bar window is a test assumption. Keep the observation time and later confirmation time separate so a chart marker does not imply advance knowledge.
• Use standard OHLC candles for these examples. Before turning recognition into a strategy, define execution timing, exits, costs and position sizing, and evaluate a later period not used to select thresholds.