A Very Brief Introduction To Candlestick Patterns: All Rights Reserved
A Very Brief Introduction To Candlestick Patterns: All Rights Reserved
Introduction to
Candlestick
Patterns
Please note that this text sometimes refers to CandleScanner software. However, the
knowledge described here is universal and may be used by anyone, no matter if
CandleScanner is used or not.
Japanese candle patterns are well known and routinely implemented in displaying price behaviour. However, when
apparent emerging patterns are analysed and discussed, it is frequently the case that the conclusions are imprecise, and,
indeed, often result in contradictory interpretations of what the patterns are actually saying. Hence, to accurately
implement a tool scanning charts for candle patterns is not a straightforward undertaking.
Candle patterns are very interesting for traders due to their simplicity, elegance and natural interpretation of market
sentiment. No matter how markets evolve, ‘patterns’ will appear on the charts. Essentially, the core premise, and
assumption, underlying the application of technical analysis is that such patterns are repetitive and detectable. Technical
analysis is widely employed in various financial markets, informing traders about the non-fundamental determinants of
price, in other words, "market sentiment."
However, observing an apparent repeating pattern on a chart can result in completely different trading outcomes.
Experienced traders know that the efficiency of an investment decision depends not only on the analytical tool(s)
employed, but also on other factors, such as risk and position management. In this regard, CandleScanner™ software
can be implemented as an objective tool aiding the trader in taking positions, based on the systematic evaluation of price
behaviour.
Trying to identify patterns manually, or visually, by merely scanning the charts can be problematic. It can be also be
dangerous, as there may be a tendency to see patterns, where there are none, in random data. For example, the very
successful trader, William Eckhardt, in the Jack Schwager book “The New Market Wizards”, says that we as human beings
don’t look at data neutrally. That is, when the human eye scans a chart it doesn’t give all the data points equal weight.
Instead, there is a tendency to focus on certain 'outstanding' cases, and to form opinions on the basis of these special
cases. That’s why when Eckhardt has an insight based on a chart pattern, he tries to reduce it to an algorithm that can be
tested on a computer. In order to investigate whether or not the insight has any value, it should be explicitly formulated,
tested and evaluated. In this regard, an objective assessment has a lower chance of being biased, which is crucial when
making trading decisions.
Finally, recalling Warren Buffet's (alleged) reflection on the markets: “I'd be a bum in the street with a tin cup if the markets were
efficient.”
For a very detailed description of each pattern, including its performance, please visit our
website at: https://www.candlescanner.com/patterns-dictionary/.
Figure 1.2. Pfizer Inc. long lines (green) and short lines (orange). Candle length depends on the current volatility of the last 25 sessions,
which means that the green candlesticks may have very different span (and still be seen as long lines). Other colors used on the chart
are explained in subsequent sections. Please note that such a color theme is implemented in CandleScanner, but the user can also
It is assumed in CandleScanner that the decisions as to whether we are dealing with a long or a short line, we take the
whole candle into account (the body with the shadows).
A common analyst’s mistake is to look at the historical chart and determine the height and dynamics of the candle, not
only on the basis of the historical prices but also on future candles. However, while the candlestick is being formed, we
do not know the future yet. Therefore, we should not use the information about the future volatility of the market.
Figure 1.2 shows that the green candles, representing a long line, have very different spans. For example, the September
long lines are significantly shorter than those as at the end of October. But, in both cases, we deal with long lines
although in different market conditions.
Long/short line is also known as a long/short day. The term "day" (short or long) refers to a single
line of candle and derives from the period, where the charts were created on the basis of daily
quotations. In the case of charts constructed on the basis of intraday prices, (for example 5
minutes, hourly or others), such a term can be confusing. For this reason, in CandleScanner the
term long/short line is used to avoid any misunderstanding.
Moreover, in Greg Morris’s book titled "Candlestick Charting Explained", such phrases as long
days and short days are referring only to the height of the candle body. This is a surprising
approach, because that would mean that the candle with a small body and with very long
shadows would be considered as constituting a short day, indicating low volatility during the
day. Therefore, it might be considered that this was some simplification on the part of Morris. To
repeat, CandleScanner takes into account the whole candle height, i.e. the body and the
shadows if they exist.
Figure 1.3. Determination of the short and the long line, depending on the average price range. The green line shows 70% of price
volatility for the last 25 bars. We can notice that when the candle height exceeds the green line it is marked as green, meaning that we
By default CandleScanner adopts the following rule determining whether the candle is a long or a short line: it sets the
current range of volatility as an exponential average distance between the highest and lowest prices of individual candles
for the last 25 candles. A candle which spans more than 70 percent of this volatility value is regarded as a long line.
Candles below this threshold are regarded as short lines.
Figure 2.4 illustrates the process of determining the short and a long line. Red lines designate the average range of
volatility for the last 25 candles. The green line in the middle indicates 70 percent of the range between red lines. Each
candle located below this threshold is classified as a short line and marked with the orange color. Dark blue and red
candles will be discussed later.
The parameter value of 70 percent was arbitrarily chosen. Normally, this depends on personal preference, and
CandleScanner allows the user to change this to meet specific requirements. Nevertheless, it is good to remember that
the more we increase this parameter, the more short lines we will have, and vice versa when we lower the value, there
will be more long lines. This also has an impact on the number of identified candle patterns found by CandleScanner.
Despite the discretion, we should remember that the optimal range is somewhere between 65 and 80 percent of the span.
The detailed construction of basic candles (and candle patterns) is described on our website:
http://www.candlescanner.com/candlestick-patterns/basic-candles/.
With high price volatility over short time intervals, when the price jumps, for example, by some 30 percent, some
candles can look as insignificant points on the chart. The use of color immediately explains and shows the nature of the
candle. In CandleScanner users can display colors of their choice in the candlestick charts using the CandleScanner Color
Theme. The following colors can be displayed:
green – long lines
orange (dark yellow) – short lines
blue – all kinds of spinning tops
red – all types of doji candles
Figure 1.4 shows a candlesticks chart using the CandleScanner Color Theme. This theme is optional, and can be switched
off if you prefer a simple candlesticks chart using just two colors. Notice that it makes to use the CandleScanner Color
Theme if you want to quickly spot the difference between short/long lines and spinning tops/doji candles. For example,
some doji candles are marked in blue color rather than red as was previously mentioned. This is because in fact such
candles are indeed spinning tops and not doji candles, but their bodies are so small that on the chart they look as if the
open and close prices are the same. Hence, most people just looking on the chart would say that they see a doji candle,
although in fact it is a spinning top.
Again, the distinction as to which exact candle we have on the chart is crucial in correctly recognizing the candle pattern.
CandleScanner enables the user to adjust the settings of the searching algorithms to meet specific requirements. For
example, the CandleScanner algorithm searching for doji candles can be adjusted to accept small deviations from the
strict doji definition which says that such candles do not have a body (open and close prices are equal). This can be
helpful while dealing with larger candles. We can then define in CandleScanner to accept as doji candles such cases
where body length is up to 3% of the whole candle length. Similarly, CandleScanner allows defining the threshold for
long/short lines.
Figure 1.5. A so-called Bearish Engulfing candle pattern is forecasting a reversal of an uptrend into a downtrend. The first Bearish
Engulfing pattern is not recognized as a valid pattern because the trend requirement was not met (candles were not above the moving
average line). In this particular case, CandleScanner required that an uptrend needs to last at least three candles (days) prior to the
occurrence of the candle pattern.
For a very detailed description of each pattern, including its performance, please visit our
website at: https://www.candlescanner.com/patterns-dictionary/.