Marubozu Candle: The Bar With No Wicks and How to Trade It

A marubozu candle is all body. Price opened, marched in one direction for the whole session, and finished on or beside its extreme — no upper wick worth measuring, no lower one either. The name is Japanese: marubozu (丸坊主) means "close-cropped head," a bar shaved bald of its shadows.

That bareness is the message. Every other candle records at least one failed excursion; this one records none. Below is the exact threshold Vike's engine uses to draw the line, the two half-strength variants it also tracks, six real detections from Binance daily charts, and the two that punished anyone who chased them.

What a marubozu candle means

Think about what a session with no wicks requires. One side was in control at the open, stayed in control through every attempted pushback, and still held the extreme at the close. A green marubozu is a session sellers never led for a single print worth recording; a red one inverts that sentence.

So the candle is a statement about conviction, not indecision — the exact opposite claim a doji makes. Where the doji says nobody won, the marubozu says nobody else even scored.

Direction comes built in. Most one-candle patterns need the trend around them to mean anything. This one carries its own sign: the engine stamps a green marubozu +100 and a red one -100 from the same rule, and the open question is never which way it points, only whether the move has anything left.

How to identify a marubozu: the exact rule

Real charts almost never print a literal zero-wick candle, so a threshold has to say how much shadow still counts as "none." Vike's line sits at five percent.

• The bar must have a real range and a real body — a flat or one-price bar is skipped.
• The upper shadow is at most 5% of the high-to-low range.
• The lower shadow passes the same 5% test.
• Green gets +100, red gets -100 — one function, sign split by colour, the same shape as the engine's engulfing detector.

How Vike computes it. batch_marubozu in vike-indicators flags a bar when upper_shadow ≤ 0.05 × range and lower_shadow ≤ 0.05 × range, with range > 0 and body > 0. Both shadows are measured against the bar's own range — nothing else enters the arithmetic.

Run that over real tape and it stays rare. Across 50,794 daily bars from 20 Binance majors the full rule fired 361 times — about one bar in 140.

The closing and opening marubozu

Vike's crate ships two half-strength relatives, and the split teaches you something about how sessions end.

A closing marubozu trims only the close side: a green bar with no upper wick, a red bar with no lower one. Price finished on its extreme, whatever happened at the start. An opening marubozu polices the open side instead — the bar launched from its extreme and never looked back, but was allowed to give some ground before the close.

Half-strength in code. batch_closing_marubozu requires only the exit-side shadow to pass the 5% test; batch_opening_marubozu requires only the entry side. A full marubozu is, by construction, a bar that satisfies both at once — and the scan confirms the identity on real tape: 361 bars pass both partial tests, exactly the full detector's count.

Frequencies on crypto are lopsided in an instructive way. My scan found 6,671 opening marubozu against 3,034 closing ones. On a market that never shuts, a bar opens mid-flow at the previous close, so launching from the extreme is cheap; still sitting on the extreme at the final print is what's expensive. When you see a closing marubozu, respect it more — the session ended at full pressure.

What the rule does not say

Two omissions separate the engine's definition from the textbook drawing, and both surprised me when I first read the source.

There is no size requirement. Nothing compares the body to the recent average — a tiny candle with trimmed wicks qualifies just as well as a monster. Every illustration you have seen draws the marubozu huge, yet size lives in different detectors entirely (batch_long_line is the one that measures bodies against their neighbourhood). The charts on this page all show large examples because I filtered for legibility, and it matters that you know that filter is mine, not the rule's.

There is also no warm-up and no context. The doji family needs ten bars of history before it can fire, because its yardstick is an average. A marubozu is judged purely against itself, so the first bar of a fresh listing can already be one — worth remembering on new pairs, where day one is often exactly this shape and means nothing beyond launch mechanics.

How to trade a marubozu

Every marubozu is a momentum reading, so the honest question is always the same: continuation or exhaustion?

In the middle of a move, lean continuation. The BNB bar from 2020-03-08 above closed at 95% body while the whole market was unravelling; four sessions later price sat 45% lower. Momentum that clean rarely dies the next morning.

Late in a move, start distrusting it. A huge full-body bar after weeks of trend is often the crowd's last, loudest push. I treat the size of the move behind it as the tell — the further price has already travelled, the more a marubozu reads like a climax than a beginning.

Use the far end of the bar as the line. A green marubozu's low is where the thesis dies: if conviction was real, price should not trade back through the whole session that announced it. That range is usually wide, which prices the signal honestly — big statement, big risk.

Do not chase the close. The candle only reveals itself at its final print, after the full move is already paid. Waiting for a pullback toward the body's midpoint cost me nothing in missed trades and saved me plenty on the entries that mean-reverted first.

Where marubozu candles fail

Both failure charts above are worth sitting with, because they fail in the pattern's own language.

XRP printed a bearish marubozu on 2024-07-07 — sellers held every print of the session — and price rose 7.1% over the next four days anyway. LINK ran the mirror image on 2026-07-26: a clean bullish bar, then a 4.2% slide. Full-session conviction told you what the crowd did that day, and the market spent the rest of the week disagreeing.

Liquidity, not psychology, drives those reversals. A bar that closes on its extreme leaves every short-term participant offside in the same direction at once, and thin books love to hunt exactly that. In my own trading the fade setup — marubozu into a prior level, then a failed follow-through bar — has been at least as productive as the continuation it advertises.

Frequency cuts the other way here. At one bar in 140 this is a genuinely rare print, far rarer than a spinning top, so it deserves attention when it appears. Rarity is not reliability, though — it only means the sample is small enough that nobody should quote you a win rate with a straight face.

Frequently asked questions

What does marubozu mean in Japanese? Marubozu (丸坊主) translates as "close-cropped" or "bald head" — the candle with its hair shaved off. The wicks are the hair; this bar has none.

Is a marubozu bullish or bearish? Whichever colour it is. Green means a fully bullish session, red a fully bearish one — the engine emits +100 or -100 from one rule. Whether the next session honours that is the real question, and the failure examples above show it often does not.

Does a marubozu have to be a big candle? Not by the engine's definition. Only the shadows are tested, each against 5% of the bar's own range. Textbooks add "long body" to the description; the detector deliberately leaves size to other rules.

How is a marubozu different from three white soldiers? Kinship, actually: three white soldiers is three consecutive sessions of near-marubozu behaviour, each policed for a small shadow on the pattern's own side. One bar of total control versus three bars of sustained control.

This is educational material, not financial advice. Candlestick patterns describe probabilities, not certainties, and every real trade carries real risk — size positions so a wrong read costs little.

See the full candlestick pattern cheat sheet