Doji Candlestick — What Indecision Really Means (and What It Doesn't Promise)

Doji is probably the most recognizable Japanese candlestick. A thin cross on the chart that every technical analysis course describes the same way: "market indecision, possible trend reversal." The problem is that when someone finally measured what happens to price after a doji — across 4.7 million candles — the result was something else entirely. It was a coin flip.

This article won't try to convince you that a doji is a magic signal. It will do exactly the opposite: show you the numbers that dismantle one of the most durable candlestick myths. Because if you're going to watch a formation, you should first know whether it predicts anything at all.

What a Doji Looks Like

Doji (from Japanese, "the same thing") is a candle where the open and close prices are practically identical. On the chart, instead of a body you see a horizontal line — a cross, a plus sign, sometimes the letter T.

Defining features:

• Open ≈ close — the body is zero or negligible; on stocks the difference is a matter of single cents, on BTC it can be a few dozen dollars against a price in the tens of thousands.
• Wicks can be anything — short, long, symmetrical or not. The extreme variants have their own names: long-legged doji, dragonfly doji, gravestone doji.
• One candle — the formation is a single line; it doesn't require a confirming candle to "count" (trading on it is another matter — more on that below).
• Context gives it a name — a doji in an uptrend is called a northern doji, and in a downtrend a southern doji. This distinction was introduced by Thomas Bulkowski so he could measure the two cases separately.

The textbook interpretation is logical and sounds nice: since buyers and sellers ended the session in a draw, the existing trend "runs out of fuel" and the odds of a reversal increase. So a northern doji is supposed to be a bearish signal, a southern doji a bullish one.

Logic, though, isn't the same thing as statistics.

What the Numbers Say

Thomas Bulkowski, author of "Encyclopedia of Candlestick Charts," tested 103 candlestick formations on a sample of 4.7 million daily candles from US stocks. For each formation he checked how often price actually does what the theory promises, and how large the move is after the breakout. Results for doji:

Northern doji (in an uptrend):

• Theory: bearish reversal. Measured: continuation of the rally 51% of the time. The formation doesn't just fail to reverse the trend — it minimally more often keeps it going. A one-percentage-point difference from a coin flip is statistical noise.
• Performance rank: 83rd out of 103 formations (1 = best).
• Frequency of occurrence: 6th out of 103 — meaning a doji is nearly everywhere. Bulkowski compares it to sand on a beach.

Southern doji (in a downtrend):

• Theory: bullish reversal. Measured: reversal 52% of the time. Here the theory does "check out" — except the edge is 2 percentage points. Bulkowski himself calls this "near random."
• Performance rank: 78th out of 103.
• Frequency: 8th out of 103 — also very common.

There is one seemingly good result: price hits the target measured by the candle's height 88–90% of the time after the breakout. Sounds great until you notice a doji has a zero body — so that "target" is often a move of a fraction of a percent. The best average 10-day move is 3.2–3.5%, which puts doji in the middle of Bulkowski's ranking. For comparison: moves above 6% are what the author considers genuinely good.

The verdict from the measurement is brutally simple: a doji doesn't predict direction. Not up, not down. It appears so often and behaves so randomly that treating it as a standalone trading signal has no statistical justification.

One important caveat: these numbers come from US stocks on the daily timeframe. On BTC or ETH — a market that runs 24/7, with a different volatility structure and no opening gaps — the distributions might look different. But the burden of proof is on whoever claims a doji suddenly starts working on crypto. Nobody has presented that proof on a comparable sample.

How NOT to Trade a Doji

Since the formation is random, the most important section of this article is the list of things not to do:

• Don't open a position "because a doji appeared." Shorting after a northern doji at a BTC top is a coin-flip bet — with commission and spread working against you. Over the long run that bet loses by definition.
• Don't treat a doji as confirmation of your thesis. That's a classic confirmation bias: you're leaning short, you see a doji, you call it a "signal." The market just said "I don't know" — and you heard "we're going down."
• Don't count on a big move after a doji. Even when you get the direction right, the average size of the move after this formation is mediocre (rank 78–83 out of 103).

What's a sensible thing to do with a doji? Treat it as information about a pause, not a forecast:

1. Check where the doji appeared. A doji at a meaningful level (support, resistance, a high-volume zone) says more than a doji in a vacuum — because the level does the work, not the candle.
2. Wait for the breakout. A close above the doji's high or below its low at least settles who won the pause. Bulkowski notes that gap confirmation after the formation improved results — on crypto, where gaps barely exist, the equivalent is a decisive breakout candle.
3. Filter by context. From Bulkowski's tests: doji near yearly lows, and doji with longer-than-typical wicks, performed better than average. That's still not an edge you can build a strategy on — but it points to the right way of thinking: candle + location, never candle alone.

A crypto example: ETH prints a doji at resistance near a local high after a week of gains. The textbook says "short." The statistics say: 50/50. A sensible approach: if the next candle closes below the doji's low and the resistance level holds — you have a setup with a defined stop location (above the high). If it closes above the high — the trend simply continues and the doji was noise. Either way, the breakout and the level made the decision, not the cross on the chart.

It's also worth understanding why the doji myth stays so alive despite the crushing numbers. First — selective memory: a doji appears so often (6th–8th place in frequency out of 103 formations) that, purely statistically, it must sometimes land exactly at a top or a bottom. Those cases end up in textbooks and on Twitter screenshots; the thousands of doji after which nothing happened end up nowhere. Second — the narrative is too nice to give up: "a balance of power before the turn" sounds like wisdom, and "noise with no predictive value" doesn't sell courses. Third — hardly anyone even knows someone measured it. Bulkowski did the work that most formation teachers never did: he checked.

There's one more practical reason to know these numbers even if you don't trade candles yourself: the crowd trades them. Seeing a doji at a top, part of the market will try to short — and just as often get stopped out, fueling the continuation. Understanding that the formation is random protects you from joining that group and lets you read its behavior from the sidelines.

Myth vs. Measurement

A doji is like a "caution" sign on the road: it tells you something might happen, but not what. Traders who treat it as an entry signal pay for that mistake spread after spread. Traders who treat it as an invitation to check the context at least aren't handing their money to randomness.

No hype: no single candle will keep your account alive. A doji certainly won't.

FAQ

Rafał — Krypto Bez Ściemy

Trader and founder of Strefa Tradingu. He’s been breaking crypto down on YouTube for years — no hype, no signals, with a focus on market structure and risk management.

kryptobezsciemy.plInstagramYouTubeKlub Zero