Long-Legged Doji: What It Means and How to Trade It

A long-legged doji is the market's loudest shrug. Price travelled a long way up, a long way down, and closed almost exactly where it opened — a near-invisible body strung between two long shadows. The Japanese canon calls it juji (十字), the cross, and the shape earns the name honestly.

Loud is the operative word. A quiet session that goes nowhere tells you nothing; a violent one that ends flat tells you both sides committed real money and cancelled each other out. Below is the exact rule Vike's engine applies, the odd discovery I made about its rickshaw man twin while reading the source, six real detections from Binance daily charts, and the two that resolved against the reader.

What a long-legged doji means

Picture the session behind the candle. Buyers drove price far above the open — far enough to leave a long upper shadow — and were thrown back. Sellers then did the same in the other direction with the same result. At the close, all of that travel had netted out to roughly zero.

That is different in kind from an ordinary quiet doji. Both close where they open; only one of them fought about it. The long legs certify that the standoff was expensive, and expensive standoffs tend to happen where trends die — which is why the pattern is worth a page even though, on its own, it points nowhere.

Across 480,125 daily bars from 376 Binance pairs, the detector fired 16,820 times — one bar in 29. Common enough to see weekly, rare enough to notice.

How to identify a long-legged doji: the exact rule

Vike's engine tests three clauses, and they use two different yardsticks — the detail most guides flatten away.

• The body is at most 10% of the 10-bar average body — judged against its neighbours, not against itself.
• The upper shadow is at least 30% of the bar's full range.
• The lower shadow clears the same 30% bar.

How Vike computes it. batch_longlegged_doji in vike-indicators flags a bar when body ≤ 0.1 × avg_body(10), upper_shadow ≥ 0.3 × range, and lower_shadow ≥ 0.3 × range, with a real range. The averaging window includes the bar itself, so the doji's own tiny body drags the average down and makes its own test slightly harder — a subtlety ported straight from the source.

Note what is missing: a body > 0 clause. A perfect doji, open equal to close to the last decimal, sails through — the PLUME chart above is exactly that, a measured 0.0% body flagged by the engine. Its cousin the spinning top explicitly skips zero bodies; this rule embraces them.

Long-legged doji vs rickshaw man

Every textbook lists the rickshaw man as a separate pattern: a long-legged doji whose body sits at the centre of the range. The crate has a separate function for it too — batch_rickshaw_man, the same three clauses plus a fourth demanding the body's midpoint lie within 25% of the range's midpoint.

Here is what I found when I measured it: over those 480,125 bars, the long-legged doji fired 16,820 times and the rickshaw man fired 16,820 times. Identical sets, bar for bar.

That is not a coincidence of my tape — it is arithmetic. The distance between body midpoint and range midpoint always equals half the difference between the two shadows, and with both shadows at 30% minimum, that distance can never exceed 20% of the range. The fourth clause tolerates 25%. It is unreachable: any bar passing the first three clauses passes the fourth for free, so the crate's rickshaw man is its long-legged doji. No chart pattern site will tell you that, because you can only learn it from the code and a half-page of algebra.

The two yardsticks, and why they matter

The body is measured against recent history, the legs against the bar itself. Each choice blocks a specific false positive.

Judging the body against the 10-bar average stops a low-volatility drift bar from qualifying — tiny is only meaningful next to what tiny has recently meant. Judging the shadows against the bar's own range keeps the "long" in long-legged honest on any timeframe and any price level: 30% of the range each way, whatever that range is.

Together they also force real geometry. Two shadows at 30% minimum leave at most 40% of the range for the body — and since the body must also be under a tenth of the average, a qualifying bar is a wide-ranging session pinched to a point. The dragonfly and gravestone doji take the opposite route: one enormous leg, one missing.

How to trade a long-legged doji

Indecision has no direction, so the candle is a flag, never an entry. Three habits keep it useful.

Require a trend to interrupt. Every example on this page printed after a move of at least 5% in the prior four sessions, because that filter is what separates a stall from wallpaper. In a sideways range I skip these entirely.

Let the next bar vote. The ETH detection above turned a decline into a 5.2% four-day recovery, and the evidence was the following session closing back through the doji's midpoint. Until that vote lands, both directions remain live.

Use the legs as the risk box. The doji's high and low bracket the entire fight between the two sides. A break beyond either end, held into a close, is the resolution — and the box is wide by construction, which is the honest cost of trading a violent stalemate.

Where long-legged dojis fail

The failure mode is specific: the trend simply keeps going, and the doji turns out to have been a rest stop rather than a top or bottom.

Both failure charts above show it. ALLO printed a textbook long-legged doji inside a rally on 2026-07-04 — and the rally ran on another 18.3% in four days. BNB did the same in August 2021, resolving 6.6% higher in the direction it was already travelling. Anyone who shorted the "exhaustion" paid for the lesson.

My own rule of thumb after watching plenty of these: the pattern marks where a reversal could start, and the market declines the invitation more often than it accepts. Four of the six charts on this page turned; that ratio came from a filtered scan, not from the raw population, and I would never quote it as a win rate.

Frequently asked questions

Is a long-legged doji bullish or bearish? Neither. It is violent indecision, and its meaning comes from the trend it interrupts plus the bar that follows. After a rally it warns of exhaustion; after a decline, of a possible floor.

What is the difference between a long-legged doji and a rickshaw man? In Vike's engine, nothing — provably. The rickshaw man adds a body-centering clause that the long-legged doji's own shadow requirements already guarantee, and 16,820 detections over 480,125 real bars confirm the two rules flag identical bars.

How is it different from a standard doji? The legs. batch_doji tests only the body; this rule adds a 30%-of-range minimum on each shadow, so every long-legged doji is a doji but only about half the dojis on crypto tape qualify as long-legged — 16,820 of 30,745 in my scan.

Can the body be exactly zero? Yes. There is no minimum-body clause, and the PLUME example on this page closed precisely on its open. The engine flags it; a spinning top detector would not.

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