Candlesticks
Japanese Candlestick Charting Techniques
Nison did not invent candlesticks. He translated them, and in doing so he handed Western traders a way to see the fight between buyers and sellers inside a single bar. Read it for the vocabulary and the psychology behind each shape. Do not read it for a tested edge, because there is not one in here.
The one idea
A bar chart draws four numbers as one thin line and treats them all the same. Nison's argument is that two of those four numbers carry almost all the emotion, and the other two are noise. The open is where every overnight rumor, every margin call, every scared short gets resolved into one price. The close is what every computer system, every margin desk, and every technician grades the day against. The distance between them, drawn as a thick block Nison calls the real body, is the session's actual verdict. The high and the low are the shadows, and the Japanese treat them as extraneous price fluctuations.
Once you draw it that way, a whole layer of information shows up that a bar chart hides. A day that opens at its low and closes at its high is a long white block, and it looks nothing like a day that covered the same range but closed where it opened. That second one is a doji, and it means the buyers and sellers fought all day to a draw. Put a doji at the end of a long rally and you are looking at buyers who ran out of conviction. Nison's fifty-plus named patterns are all variations on this: the shape of the body, and where it sits relative to the body before it.
The second half of the book carries the part most readers skip. Nison does not trade off candles alone. He uses them as a trigger on top of ordinary Western analysis: a trendline, a 50 percent retracement, a Wyckoff spring, an old support level that flipped to resistance. He calls this the rule of multiple technical techniques, borrowed from Arthur Sklarew. A hammer sitting on nothing is a curiosity. A hammer sitting exactly on the level where a broken support turned into new support is a trade. That is his whole method, and he is explicit that it is his whole method.
What it actually teaches
Nison's own trading sequence, reconstructed from how he actually places trades in Chapter 20 rather than from the order the patterns are taught.
READ THE BODY, NOT THE WICK
The thick block between open and close is the signal. White (or green) means the close beat the open. Black (or red) means it did not. A candle with no upper shadow is a shaven head, one with no lower shadow is a shaven bottom. Nison is blunt that the Japanese treat the shadows as extraneous. This matters because most of his pattern rules compare real body to real body and ignore the wicks entirely. The bearish engulfing pattern requires the second body to swallow the first body, and Nison says explicitly it need not engulf the shadows.
NAME THE TREND BEFORE YOU NAME THE PATTERN
The same shape is bullish or bearish depending only on what came before it. A small body at the top of the range with a lower shadow twice the height of the body is a hammer after a decline and a hanging man after a rally. Identical line, opposite meaning. Nison's rule is harder than most people apply it: place a new position on a reversal signal only if that signal points in the direction of the major trend. In a bull market a top reversal is a signal to sell your longs, not to go short.
MATCH IT TO THE ACTUAL CRITERIA
The rules have numbers. Hammer and hanging man: body at the upper end of the range, lower shadow at least twice the height of the body, no or very short upper shadow. Dark-cloud cover: the black candle opens above the prior day's high, then closes more than halfway down the prior white body. Piercing pattern: the white candle opens below the prior day's low and closes above the midpoint of the prior black body. Rising three methods: a long white candle, then about three small falling bodies that stay inside its range, then a white candle closing above the first candle's close. Three black crows: three declining black candles, each opening inside the prior body, each closing at or near its low.
WAIT FOR THE NEXT SESSION TO CONFIRM
Nison is insistent about this for the weaker signals and it is the step readers drop. The hanging man requires bearish confirmation: the more the next day opens below the hanging man's body, the more likely it was a top. The inverted hammer requires the next session to open above its body. A doji star in an uptrend is voided if the next candle is white and gaps higher. Doji themselves are asymmetric: Nison says they call tops well but lose reversal potential in downtrends, so a doji needs more confirmation to mark a bottom than a top.
STACK IT ON A WESTERN LEVEL
This is the rule of multiple technical techniques and it is where Nison thinks the real power is. His worked example: in late 1989 gold traded near $380 and he called resistance at $425 to $433 from four independent readings. A 50 percent retracement of the 1987 high at $502 and the 1989 low at $357 gave $430. A double-bottom measured move gave $425. The late 1988 high was $433. An Elliott fourth-wave count put the ceiling at $425. Gold peaked at $425. No candlesticks were used in that call at all. Candles are what he adds on top to time the entry.
BORROW THE STOP AND TARGET FROM SOMEWHERE ELSE
Candles do not give you either one. In his own corn trade he buys at $2.68 on a hammer confirmed by a higher open, but the stop at $2.64 comes from a support line and the $2.79 target comes from a downward sloping resistance line. In coffee he shorts at $.9015 with the stop above a window at $.9175 and a target of $.8675 from prior-year support. The window (his word for a gap) is doing the work: Japanese practice says corrections go back to the window, so an open gap is your resistance until it closes and the buying continues past it.
What it looks like Monday morning
Switch your charts to candles if you have not, then stop looking for pattern names and start reading two things: how big the body is relative to recent bodies, and where the body sits inside the day's range. That alone changes what you see. A doji after three long white candles at the top of a run is the market telling you the buyers just lost the argument, and no bar chart shows you that.
Then do the thing nobody does with this book, which is take the second half seriously. Mark your levels first: yesterday's high, the old support that broke, the 50 percent retracement, the gap that has not filled. Only look for a candle signal when price arrives at one of those. A hammer in the middle of nowhere is not a trade and Nison would not take it either. And write your stop and target from the level, never from the candle, because the candle does not have one.
| Pattern | Where the white candle closes | Signal |
|---|---|---|
| On-neck | Near the low of the prior black candle | Bearish. Sell if price breaks the white candle's low |
| In-neck | Slightly into the prior black body | Bearish |
| Thrusting | Well into the black body but below its midpoint | Bearish in a falling market, bullish continuation in a rising one |
| Piercing | Above the midpoint of the prior black body | Bullish reversal |
Nison allows slack on the dark-cloud cover's 50 percent penetration but almost none here, because getting the depth wrong flips a buy signal into a sell signal.
A trend reversal signal implies that the prior trend is likely to change, but not necessarily reverse.Steve Nison, Japanese Candlestick Charting Techniques
Where it fails
- There is not one statistic in the book. Nison names over fifty patterns and tests none of them. There is no win rate, no sample size, no holding period, no comparison against a coin flip. The entire evidence base is hand-picked chart exhibits from 1987 to 1990, mostly pit-traded futures, where the pattern is labeled at the turn after the fact. When a pattern fails, as hammer 4 does on his lumber chart, he explains afterward why you should have skipped that one (the prior day was a long black shaven-head candle, and the hammer broke the January 24 support). A rule that comes with an unlimited supply of after-the-fact exceptions cannot be wrong, which means it cannot be tested. The large-sample work that came later, notably Marshall, Young and Rose on Dow stocks from 1992 to 2002 and Bulkowski's pattern encyclopedia, found most of these patterns land close to a coin flip, and several so-called reversal patterns resolve as continuations more often than not.
- The edge he sold was informational, and he personally destroyed it. Read his own pitch: candlestick techniques are for the most part unused in the United States, and some patterns may allow you to get the jump on those who use traditional Western charting techniques. He lists seven vendors that offered candle charts in 1991. His bold prediction was that technicians would abandon bar charts entirely. He was right. Candles are now the default chart on every broker and every free platform, and TradingView, thinkorswim and every scanner on the internet flag hammers, engulfings and dojis automatically. Whatever advantage came from seeing a reversal a Western reversal day would have missed is gone by construction. What is left is a signal so widely visible that it is at least as likely to be fade material as follow material.
- The definitions bend until they cannot fail. Nison admits he found conflicting Japanese definitions and chose, in his words, the rules that increased the probability that the pattern's forecast would be correct. That is fitting the rule to the outcome with no out-of-sample check. Then the chart examples bend the fitted rules anyway. A crude oil candle that misses engulfing by 8 ticks is counted as a bearish engulfing pattern. A hammer whose lower shadow is not twice the body is counted. A doji star whose body does not gap above the prior body is counted. A morning star whose third candle is not long is counted. His own defense is that candlesticks have guidelines, not rigid rules. Fine, but then two readers looking at the same chart will not see the same patterns, and no backtest of yours will match his.
- Half the patterns need a real opening gap, and modern markets do not give you one. Dark-cloud cover needs the second day to open above the prior day's high. Piercing needs it to open below the prior day's low. Stars need the small body to gap clear of the body before it. Upside-gap two crows needs two gaps. Nison already saw the problem in 1991 and wrote that stocks often open at, or very near, the prior session's close, so you should allow more flexibility with candlestick indicators with equities. Since then it got worse, not better. Twenty-four-hour markets in spot FX and crypto barely have a meaningful open at all, and in liquid equities most of the gap has already been traded away in the pre-market. The abandoned baby, which he calls very rare in 1991 futures, is a museum piece now.
- It is half a system, and he says so. There are no price targets, no stop rules, no position sizing, no risk framework anywhere in the book. Every complete trade in Chapter 20 gets its structure from Western tools: a head and shoulders neckline, a Wyckoff spring, a change-of-polarity level, a stochastic divergence, a prior-year support price. The candles only decide when to pull the trigger. A reader who takes the first eight chapters and skips the rest, which is what most people do, walks away with a naming system and no way to size, stop, or exit a position. And the trigger itself requires the close, so you either need a market-on-close order or you have to guess the close a few minutes early, which is real execution friction the book waves away.
- The doji advice broke when timeframes shrank. Nison warns that doji only mean something in markets where there are not many doji, and that candlestick analysis usually should not use intra-day charts of less than 30 minutes because below that most lines become doji or near-doji. The default retail chart today is a 1-minute or 5-minute chart, and the modern tape produces small real bodies constantly. The same erosion hits spinning tops, harami and tweezers, all of which depend on a small body being unusual. Apply his rules to a 5-minute chart and you will find a reversal signal every few bars, which is the same as finding none.
- A large chunk of the page count is filler he tells you to read elsewhere. Chapters 11 through 19 introduce trendlines, retracements, moving averages, RSI, stochastics, momentum, volume, Elliott Wave, Market Profile and options. Nison calls these introductions broad, and admittedly cursory, and points you to John Murphy's book instead. He is right to. The Market Profile and tick-volume chapters are tied to early-90s CQG and Chicago Board of Trade products and read as period pieces. If you already know Western technicals, roughly a third of the book is a thin re-tread; if you do not, it is too thin to learn from.
Who it is for
Buy it if
You already have a system that tells you where to act and you want a better way to read what price is doing when it gets there. This is the reference for that: a clean, well-organized taxonomy of what each candle shape means about the fight between buyers and sellers, written by the person who introduced it to the West. Chapters 3 through 8 are the whole value, and they read fast.
Skip it if
You want a tested edge, a trading system, or numbers you can backtest. There are none here, and the ones you construct from the book will not survive contact with a spreadsheet. Skip it too if you trade crypto or spot FX, where the opening gaps that half these patterns require do not exist, or if your working timeframe is under 30 minutes, which Nison himself says is too fast for this to mean anything.
As an Amazon Associate I earn from qualifying purchases. It costs you nothing extra, and it does not change what gets recommended or what the verdict says.
Reviewed here is the original 1991 text. The link goes to the Second Edition, which is the one still in print and adds later chapters.
Back to all titles