Friday, September 5, 2014

ChartDNA Patterns Matched With Technical Indicators

ChartDNA does an excellent job of matching and analyzing patterns against your watchlist of stocks; identifying which patterns are best traded long and which may be better used on the short side.

However, ChartDNA also offers the facility to look at supporting technicals. Technicals which may also help in the filtering of patterns to trade.

As an example, we take a look at one 6-character pattern and compare it against some key technical indicators.  The pattern under review is #YBBHBB.  The presence of four 'B's make it a 'bullish' pattern sequence, although as of September 5th 2014, it has only a 27% success rate as a long trade using a 15% target and 5% stop.


Are there any technical markers which we can use to better select longside opportunities with this pattern?

As a starter, I selected some key technical indicators and marked their position at the time of the pattern trigger.

The indicators selected were:

  • 50-day Simple Moving Average
  • 200-day Simple Moving Average
  • Directional Movement System [14 period]
  • Aroon Up/Down [20 period]
  • Relative Strength Index [14 period]
  • Slow Stochastics [14 period]
  • MACD [12,26,9]

Next, I took a look at the values of the indicators for different returns. First, it was the indicator ratings of the 5% losers, then ratings of matches which offered gains of 15% up to 30%. It should be noted, there were only a low number of matches for the 20% gain and higher, so it's hard to draw conclusions for these returns. 

Each of the charts to follow plot the 95% confidence range for a given indicator, as classified by return, with a mean value. 

Simple Moving Averages

Given the presence of four bullish candles within the six character sequence it was no surprise to see this pattern occur when the faster 50-day SMA was trending above the 200-day SMA: the 50-day SMA was on average 4.8% above the 200-day MA. However, when looking at the differential between these moving averages at the time of the trigger there was little on offer between winners and losers. 


Directional Movement System

The Directional Movement System offers three core variables: a bullish trend measure (+DI), a bearish trend measure (-DI) and a trend strength measure (ADX). 

Again, because of bullish nature of the pattern, the +DI and -DI values fell in a relatively narrow spread at the time of the trigger. 

In the case of +DI, only the 15% gain had a significant association with a +DI value above 30. However, larger gains were associated with a falling +DI, with an average value below matches which lead to a 5% loss. With the low number of matches it's difficult to deduce if the apparent contradiction has merit (i.e. profitable matches also occurred for +DI below 25). 


The -DI was a little clearer in that profitable matches, at all % gain levels, occurred when -DI was below 16 at the time of the trigger.


Trend strength (ADX) may also be a factor for profitable matches. It was interesting the ADX 95% confidence range was above 20 for both winning and losing matches (An ADX above or below 20 is typically used to differentiate between a trending or range bound market, which in itself is a useful lead to help determine which indicators to use when trading). This suggests #YBBHBB is a continuation pattern. 

In profitable matches, the prevailing trend strength was stronger. Ideally, the ADX should be above 26 for a profitable match. 


Aroon

Like the Directional Movement System, the Aroon is another trend-centric indicator. And like +DI/-DI, it gave a contradictory output.  In the case of Aroon Up, a 15% gain often had a value above 80, but for higher returns the 95% confidence bands widened, and the Aroon Up values sank...


On the flip-side, the higher the gain, the higher Aroon Down was too.  Athough, there was less of a relationship between Return and Aroon Down. 


Relative Strength Index

Relative Strength is a momentum indicator and matches were found to occur within a relatively narrow band, even for the low number of matches at the higher return.  However, the same contradiction appeared in the RSI as it did for the trend indicators, Aroon and ADX.  The 15% gain appeared to be associated with an RSI above 62, but when higher gains were looked at, the RSI fell to values associated with matched losing trades. 


Slow Stochastics

The second momentum indicator looked at was Stochastics. Both %K and %D values were analyzed.  The %K values appeared to be a little more relevant than %D. And like RSI, values fell within a narrow range for winners and losers. 

Winning trades tended to be associated with a %K above 80 at time of the match, with %D in the 80s also helpful when looking to identify potential winners. 


MACD

The Moving Average Convergence Divergence proved to be less helpful when looking for winning matches.  The 5% losers in 78% of cases had a MACD above the Signal line, compared to 81% for the 15% gainers.

Summary

While pattern #YBBHBB is not one of the most successful patterns available, its odds for success can be improved by looking for certain technical conditions. Key factors to look for are:
  • %K above 80
  • ADX above 26 (or 28)
  • -DI below 16
  • +DI above 30

Out-of-Sample Matches

The test data set only included matches which hit target or stop.  However, there are a number of open matches which have yet to reach their target (or stop).  

Applying the above four filters to the 49 matches currently open, reduces the number of potential candidates to 5 (current return listed).

Conoco Phillips: May 22nd 2014: +2.03%  Target: $90.00  Stop: $74.35
Conoco Phillips: June 11th 2014: -2.02%   Target: $93.73  Stop: $77.43
Colgate Palmolive: March 18th 2014: -0.33%  Target: $74.20  Stop: $61.29
Devon Energy: June 2nd 2014: -2.09%    Target: $84.93  Stop: $70.16
Apache Corp: June 6th 2014: +6.37%    Target: $106.78   Stop: $88.21


Conclusion

Different ChartDNA patterns will exhibit different technical strengths and weaknesses which further analysis can elucidate. For example, pattern #YBBHBB was responsive to trend strength (ADX), but another with characteristics more suited to a reversal pattern will unlikely be so influenced. 

A greater understanding of the relationships between technical indicators and patterns can improve trade performance - allowing you get the most out of ChartDNA. 

No comments:

Post a Comment