Doji Candles
— Telling Standard, Gravestone, Dragonfly, and Long-Legged Apart
A doji is a candle whose open and close are nearly identical, so the body all but disappears. There isn't just one shape — there are four, and even the same-looking doji reads differently depending on where it shows up. We start with how to tell the four shapes apart.
What It Means for the Body to Disappear
A candle's body is the distance between the open and the close. When that distance is close to 0, it means that after a full day of buying and selling, the price ended up back where it started. No matter how far it moved up or down along the way, the conclusion left standing is "a draw."
That's why a doji doesn't point in a direction on its own. But if there was a strong move in one direction leading up to it, the doji becomes a record that that move just met resistance for the first time. The body/wick concepts from Candlestick Chart Basics apply directly here.
The Four Shapes and Where the Wick Sits
There's really only one thing that separates the four types of doji: which side the wick is on, and how long it is.
| Shape | Appearance | What happened intraday |
|---|---|---|
| Standard Doji | A cross (+). Upper and lower wicks similar | Pushed both ways, but neither side could hold ground |
| Dragonfly Doji | A "T" shape. Only a lower wick | Sold off hard, then fully recovered by the close |
| Gravestone Doji | An upside-down "T" (┴). Only an upper wick | Rallied hard, then gave it all back by the close |
| Long-Legged Doji | Both upper and lower wicks very long | Wide range intraday, but no direction ever took hold |
← Scroll sideways to see the full table.
"Dragonfly" and "gravestone" are unfamiliar names, but the structure behind them is simple. Whichever side has the wick is the direction the price got pushed to, then pulled back from. A long lower wick is a record that buyers stepped in and supported the price from below; a long upper wick is a record that sellers pushed the price back down from above.
The Same Doji Means Something Different Depending on Where It Shows Up
This is where doji interpretation goes wrong most often. Two identically shaped cross candles can read completely differently depending on what was happening right before them.
- After a long rally — a record that the upward push just met balance for the first time
- After a long decline — a record that the downward pressure just found support for the first time
- In the middle of a sideways range — no special information. It was an undecided stretch to begin with
- On lower-than-usual volume — the body may simply not have formed because few participants showed up
Reading meaning into a doji that shows up in the middle of a sideways range is the single most common overinterpretation beginners make. A doji is the shape you get when there's no conclusion to begin with, so in a stretch that already had no conclusion, it isn't telling you anything new.
What to Check Next — What Comes After a Doji
- Pin down the prior trend first. A moving average or a trendline works well as the reference
- Check the volume. A doji on above-average volume carries more weight as a record than one on below-average volume
- Wait for the next candle. A doji is a question, and the answer is in the candle that follows
- Check the price zone. A doji that shows up right at a support or resistance boundary has much clearer context than one that shows up in open air
Common Misunderstandings
"The body has to be exactly 0 to count as a doji" — not true. In practice, a body within roughly 5–10% of that candle's total range is usually treated as a doji. The exact threshold varies by charting tool, so it's worth checking how your own platform defines it.
"A doji means the price reverses" — a doji is a record of balance, not an announcement of direction. It's common for a doji to appear in the middle of a strong trend and for the price to simply keep going the same way afterward. The most accurate thing you can say after seeing a doji is "we don't know yet."
Push one step further and you run into another question — "how do I actually check whether this pattern works?" There are so many candle patterns that if you test a bunch of them at once, at least one will look like it works just by chance. The way to filter out that illusion is multiple-testing correction, and the trap of building a rule that only fits past data is overfitting.