Doji Candlestick Types: Five Shapes, One Number

By Uladzimir Leuchankou, founder of Signal2Forex · 6 November 2026

What is a doji candle? Just a candle where the open and the close are the same price — a small cross on the chart? I suppose that is an answer too. But for me a doji is more than a shape. It is a number — the size of the body against the size of the whole candle.

The short answer comes first. There are five doji candlestick types — standard, long-legged, dragonfly, gravestone and four-price. All five say one thing: buyers and sellers ended the session where they started. So the doji candle meaning is a pause. The direction comes later, from the next candle.

I took every daily candle of GBP/USD from 2 January to 27 October 2026 — 213 trading days — and looked for each type. All five were there. The dates in the table are real, and you can open the same days in your own terminal.

Five doji candlestick types

Most pages draw the types of doji as perfect pictures. A live chart is not so tidy. So I describe each type with its real day. The wicks are in pips, and one pip on this pair is 0.0001.

The first type — the standard doji. Some traders call it the classical doji or the common doji. On the chart it looks like a plus sign — the doji cross. On 22 July GBP/USD opened at 1.3375 and closed at 1.3375. The upper wick was 20 pips and the lower wick was 20 pips. This is the cleanest standard doji candlestick of the year. It tells you that neither side won that day.

The second type — the long-legged doji. The body is the same thin line, but both wicks are long. On 29 January the pair moved 105 pips between the high and the low and closed three pips from the open. A usual day this year was 72 pips. That day the price went far in both directions and came back — a busy day with no result.

The third type — the dragonfly doji. The open and the close sit at the top of the candle, and the long wick points down. Sellers pushed the price lower during the day, and buyers brought it back by the close. The closest day I found was 27 August — a lower wick of 22 pips and an upper wick of six. This is not a perfect letter T. Perfect ones are rare on a big pair.

The fourth type — the gravestone doji. People often search for the opposite of dragonfly doji, and this is the one. The open and the close sit at the bottom, and the long wick points up. On 15 June the upper wick was 49 pips and the lower wick was two. Buyers lifted the price by almost half a cent and gave all of it back.

The fifth type — the four-price doji. Open, high, low and close are one price, so the candle is a flat line. Does it happen on a daily chart of a big pair? In my table there is one such row — 1 January, a holiday, with 1.3448 in all four columns. Nobody traded. On a one-minute chart at night you will see these lines often, and they say the same thing — no trades, no information.

A number instead of a shape

I write trading robots, and a robot cannot look at a picture. It needs a rule with a number in it. The rule “open equals close” sounds exact, but it almost never works. In 213 days the open and the close matched to the pip only twice — on 22 May and on 22 July.

So chart platforms use a limit. The body needs to be smaller than some share of the range — the distance from the high to the low. A common limit is 5%, and some scripts use 10%. Take 28 July. The body was one pip and the range was 37 pips. One divided by 37 is 2.7% — a doji under both limits.

How many days pass? With the 5% limit — nine days out of 213, about one day in 24. With the 10% limit — 21 days, about one in ten. This is one parameter, and it changes the number of signals more than two times. If you test doji candlestick patterns, you need to write down which limit you used. Without it two traders talk about two different sets of candles.

There is one more use for the limit. It separates the doji types from their neighbours. A candle with a body of 15% and two wicks is a spinning top. A small body at the top with a long lower wick is a hammer — I wrote about such long wick candles on another page. The shapes look close. The numbers are different.

Two dojis in one week

In the middle of July the pound was falling. From the open of 16 July to the close of 21 July it lost 165 pips. Then came 22 July — the clean doji from the table. Most pages about types of doji candlestick patterns say the same thing here: a doji after a fall warns of a turn up.

The price did not turn. On 23 July the pair opened at 1.3375 and closed at 1.3315 — 60 pips lower. This is about $600 on one standard lot of 100,000 units, and $6 on a 0.01 lot. A trader who bought on the doji alone lost money the next day.

Four trading days later, on 28 July, there was a second doji. The body was one pip. On 29 July the pair closed 80 pips higher, and on 31 July it closed at 1.3482. This time the turn was real.

A one-candle doji, the same pair, the same week — and two opposite results.

The count after the doji

One week proves nothing, so I counted all 21 doji days under the 10% limit. For each day I asked two questions. Where was the price going in the three days before? Where did it go on the day after?

Eleven times the next day went against the earlier move. Nine times it continued the move. Once the next day closed flat. Eleven against nine is close to half and half.

Here I do not agree with most pages about doji candlestick types. They call the dragonfly a buy signal and the gravestone a sell signal.

Thomas Bulkowski tested these candles on stock charts for his Encyclopedia of Candlestick Charts. The dragonfly, in his words, “is a reversal candle, but only half the time” — 50% in his test. The gravestone worked as a turn down 51% of the time. He calls such results random, and my small count on the pound says the same.

My own two days went against the picture in the book. After the gravestone of 15 June the pair rose 14 pips. After the dragonfly of 27 August it fell 58 pips.

There is another detail. Steve Nison brought these candles to Western traders in his book Japanese Candlestick Charting Techniques. He wrote that a doji matters more after a long rise than after a fall. The long-legged doji of 29 January fits this. It came two days after the high of the year, 1.3867, and the next day the pair fell 125 pips. But one day is one day.

I do not know whether the count is the same on the one-hour chart. I counted only days, only one pair and only ten months. This is a small sample, and I would not build a system on it.

Doji as a risk measure

So is the doji pattern useless? For me it is not. A doji does not tell me the direction, but it gives me two exact prices — its high and its low. This is a ready measure of risk, and usually a small one.

The usual plan is simple. You wait for the next candle. If the price goes above the doji high, you buy, and the stop goes under the doji low. If the price goes below the low, you sell, and the stop goes above the high.

On 22 July this meant a sell at 1.3354 and a stop at 1.3396 — 42 pips. With a 0.10 lot this stop costs $42, or 0.42% of a $10,000 account. The same share of a $100 cent account is 42 cents.

I tested this plan on the same 21 days. The entry was one pip behind the doji, and the exit was at the close of the next day. Sixteen times the next day broke only one end, and nine of these trades closed in profit. Four times the price broke both ends, and the first order was stopped. Once nothing was broken. The total for 20 trades was plus 18 pips before the spread. After the spread nothing is left.

This is why I do not use the break of a doji as a signal. It is only a clean place for an order. The reason for the trade needs to come from something else — the trend on a higher chart, a level the price has already tested, or the rules of a tested robot.

On my live account — you can watch it on our YouTube stream — the Forex V robots work on the one-hour chart of 14 pairs. Their rules are built on indicators, and every rule is a number. If you trade by hand, you need the same order — first the reason, then the doji as the place for the stop.

Green doji and small charts

Two small questions come up often. The first — what does a green doji mean? The colour only shows whether the close was a little above or a little below the open. On 28 July the close was one pip higher, so most terminals paint this candle green. One pip is not information. A green doji and a red doji are the same candle.

The second question — the timeframe. On a one-minute or five-minute chart doji candles appear all the time, because in a quiet minute the price simply does not move. The spread there is often bigger than the candle. I look for a doji on the four-hour and the daily chart, where one candle holds the work of many thousands of orders.

Five shapes, one limit for the body, the move before the candle, the candle after it, the stop behind the wick — all this is one small system, and the doji is only one part of it. The more of these parts you write down as numbers, the steadier your trading becomes.