September 2004

The February 2004 issue of 'Technical Analysis of Stocks and Commodities' magazine contains an article by Dan Valcu titled 'Using The Heikin-Ashi Technique'.  Too often traders hear about a technique and think the 'holy grail' train is leaving the station and they rush to get on board without taking time to understand what it is all about.  The purpose of this article is to comment in greater detail on the visual presentation created by the mathematics of the method.

Mr. Valcu says that 'heikin' in Japanese means 'average' and 'ashi' means 'bar'.  So a literal translation would be 'average bar'.   Indeed, the method employs an averaging technique as follows:

• haClose = (Open + High + Low  + Close) / 4
• haOpen = (haOpen(previous bar) + haClose(previous bar))/2
• haHigh  =  Maximum(High, haOpen)
• haLow  =  Minimum(Low, haOpen)

Now for those who have pulled out the Valcu article and compared his formulas with those given above, please do not be too quick to claim that I made a mistake in plagiarizing the formulas.  My formulas are equivalent and it represents one of the criticisms I have.

haHigh and haLow

Mr. Valcu's formulas in the article were give as:

• haHigh  =  Maximum(High, haOpen, haClose)
• haLow  =  Minimum(Low, haOpen, haClose)

It is mathematically impossible for the haClose to be higher than the bar High, or lower than the bar Low.   haClose is an average of the bar's open, high, low and close.   The open must be in the high-low range.  The close must be in the high-low range.   The low must be equal to or lower than the high.  Therefore, the haClose can never be higher than the High, nor lower than the Low.

Because the haClose can never be higher than the High, the Heikin-Ashi High does not need to test for the haClose as a possible price that would set haHigh.   Choosing the higher of High and haOpen is sufficient.   The same reasoning applies to picking a price for the Heikin-Ashi Low.   Choosing the lower of Low and haOpen is sufficient.  haLow does not need to consider haClose because haClose will never be lower than the Low.

I consider it unfortunate that Mr. Valcu did not understand these principles before he published his article.  And, every programmer who published script code to implement Heikin-Ashi in their charting package used the Valcu formulas with scripts similar to this example:

• haHigh = MaxList( H, haOpen, haClose);
• haLow = MinList( L, haOpen, haClose);

Not one of the twelve programmers who published scripts in Stocks and Commodities magazine pointed out that testing for haClose is unnecessary because it is an impossibility.    It does not hurt to test for it, but it is an unnecessary step.  Missing something obvious like this makes me wonder just how much serious thinking is being made to understand what this technique is all about.   Now, let's leave that issue and continue with the analysis.

haClose

The Heikin-Ashi Close is the average of four bar prices:  open, high, low and close.   This creates an interesting effect in strongly trending markets which I feel is misleading for chart readers.  Let me illustrate the effect with the following example.

The example shows the original bar data in the top half of the chart, and the Heikin-Ashi method in the bottom half.  Ensign Windows was used to prepare the examples.  Bars 1 through 4 are strongly trending up, and bars 5 through 8 are strongly trending down.  Now permit me to point out several things by comparing the two images.

The Heikin-Ashi data points are also shown on the original chart using small red dots, connected by solid red lines through the highs and lows, and a dotted red line through the closes.   These dots and lines will aid in the comparison of what Heikin-Ashi is doing to 'average' the original bar data.

In an Up candle the haClose will always be below the actual close, and in a Down candle, the haClose will always be above the actual close.  These two principles are illustrated by comparing the position of the close red dots to the bar closes in the Original chart image.  In fact, haUp candles will ALWAYS have a high wick, and haDown candles will ALWAYS have a low wick.  This is a built in behavior that may surprise most Heikin-Ashi candle readers.  It is one of the primary areas I feel is misleading.

Note:  haUp candles may or may not have a low wick.  haDown candles may or may not have a high wick.

A wick on the top of a regular Up candle implies that selling pressure has moved the market back down from the high.  Thus, I consider it misleading to see a high wick on a Heikin-Ashi up candle when no selling pressure is present.   The inverse applies to low wicks.   A wick on the bottom of a regular Down candle implies that buying pressure has moved the market off of the low.  Again, it is misleading using conventional interpretation for low wicks to be present on a Heikin-Ashi down candle when no buying pressure is present.

Mathematically the haClose can never exceed 75% of the original bar's range.  75% would be achieved when the Open and the Close occur at the extreme of the bar's High.   In that case, haClose = (H+H+H+L)/4.    Simple example:  O=4, H=4, C=4, L=0,  so haClose = 12 / 4 = 3    So the maximum haClose value is 3/4th of the range because the range was 4.   Thus the high wick size in an Up candle will be 25% of the original bar range or greater.   The low wick size in a Down candle will be 25% of the original bar range or more.

In the Chartpoint Magazine, No. 12 (2003), Yashuji Yamanaka gives five rules for trading the Heikin-Ashi charts.  His Rule 2 reads, 'Positive candle with upper shadow means "strong BUY"', and 'Negative candle with lower shadow means "strong SELL"'.  I have proved out that every haUp candle must have a high wick, and every haDown candle must have a low wick.  Therefore, Rule 2 would have EVERY Heikin-Ashi candle be either a 'strong BUY' or a 'strong SELL'.  This obviously is not the case, so I must conclude that Yamanaka's Rule 2 is an illogical statement.

haOpen

The haOpen formula can be stated more simply as the midpoint of the prior Heikin-Ashi bar's candle body.  See the graphical illustration of this where the cyan lines from the prior bar's candle body range point to the candle body midpoint.  This midpoint is used as the open of the following Heikin-Ashi bar.

The haOpen can be outside of the original bar's range.  Therefore, the range of the Heikin-Ashi bar is extended to include the haOpen price.  This extension is done by choosing the higher of High and haOpen for the haHigh, and the lower of Low and haOpen for the haLow.  Mr. Valcu describes this process as eliminating 'irregularities from a normal chart', and creating a 'better picture of trends'.  My opinion is that this process is creating misleading perceptions.  Let's look again at the example.

One misperception in the Heikin-Ashi chart is the absence of gaps.   There are 6 gaps in the original chart and they have all been 'averaged' out of the picture.  If gaps mean something to you either as an indication of momentum or a price level that will eventually be filled, you will have to do without that insight when you use Heikin-Ashi charts.

Another misperception in the Heikin-Ashi chart is the length of the bars.  In our example many of the HA bars are twice as tall as the original bars.  HA bars will always overlap a portion of the bar on its left-hand side.  In the up trending portion of the example, the HA Lows are all lower then the original lows, giving the impression the market traded at prices during that time period when no such trading occurred.   As an example, consider bar #3.  The original bar price range is from 700 to 740.   The Heikin-Ashi bar implied that during the #3 time period, the trading was from 640 to 740.  That is misleading.  The visual presentation does not make any differentiation between the portion of the range that is actual and the portion that is invented.

Another misperception is the combination of bar #4 and bar #5.  On the original chart, these two bars make a formation known as a Key Reversal Pair.  That significant information is totally lost in the Heikin-Ashi chart.  In fact, bar #5 on the HA chart is shown as an Up candle which is 100% opposite what actually happened.  That too is misleading in my opinion.

Summary

I guess by now you have concluded I am not overly impressed with the Heikin-Ashi method.  It may be serving a beneficial purpose for many of you, and if so, that is wonderful.  I encourage you to continue using what works for you.  Heikin-Ashi charts are included in Ensign Windows because users asked for it.  But, I do not know if it is going to help anyone trade more profitably.  Seasoned trader Ira Tunik recently stated,  'There are those that are constantly looking for the Holy Grail and [think] every new or revived study or tool is necessary.  Over the years I have found that the majority of the exotic, complicated and supposedly new studies don't help anyone's trading ability or profitability.'

Whatever the case may be, at least by reading and understanding the points made in this article you will be using the Heikin-Ashi method better informed about how it is creating 'average bars'.

Recently a new indicator was added to Ensign Window’s study list.  It is called the Aroon indicator.  The indicator is supposed to allow you to anticipate changes in price from trending to trading range.  At first glace the indictor looks very complicated and seems to make no sense.  With a little time, education, and experience I have found it to be most enlightening.

The description of how the indicator works is fascinating stuff.  It 'measures the number of periods that have passed since the most recent x-period high and x-period low.  Therefore, the Aroon indicator consists of two plots; one measuring the number of periods since the most recent x-period high (Aroon Up) and the other measuring the number of periods since the most recent x-period low (Aroon Down).'  Now I am almost sure that most of the people reading this article are just eagerly awaiting a further discussion of how the indicator is plotted in a Stochastic like scale etc, etc, etc.  Yeah sure.  Just in case you are really interested, a more complete discussion of the indicator can be found on the Internet using a Google search.

We are traders, and as such it is extremely important to know how to use indicators to our advantage.  I have never felt the need to be thoroughly educated on how the thing works.  Just give it to me and let me play with it long enough and I will either make it work or throw it away because it is impossible to figure out.  Right?  Ok, I have been playing with the Aroon for a few days and I think I have figured out a way to use it which I will share with you.

As you can see below the Aroon consists of two lines.  One red (the Up line), and one blue (the Down line).  In the system description there is a discussion about  three thresholds.  Those being the 70, 50, and 30 thresholds.  I have discovered that if I pay attention primarily to the red line’s relationship to these thresholds I can determine what the market is telling me about trend.  It also begins to become clear that I can use this indicator to trade.  In fact I can almost make a complete system out of the indicator. 

Rule #1:   If the red up line goes to the top of the range (Window) there is a possibility of an up trend beginning.  If the red up line stays generally between the top of the window (100) and the 70 threshold marked by the orange line on the example chart, the up trend is in motion.  The closer it remains to the top the stronger the trend.

Rule #2:   If the red up line retraces to and breaks through the 50% line (red dashed line) the trend is in serious trouble.  If it reaches the bottom of the scale/window, the trend is usually dead and a down trend may be beginning.

Rule #3:   If the red and blue lines separate and remain separated the trend is strong and I aggressively add contracts on price pull backs.

So, in summary, what we have is a strong trend if the red line goes to the top and stays there, a failing trend if it pulls back to the 50% threshold and a possible reversal into a down trend if it falls below the 30% threshold.  The reverse is true for a down trend.  If the red line breaks up through the 50% line the down trend is weakening, etc.

So how do we trade this thing?  I have discovered that if I merely watch the relationship between the red and blue lines I can use their crossovers for entry and exit signals.  Look at the chart below which I have marked up with arrows for the entries to trades.  This is real data for Sep 16^th and the trades worked as advertised.  The day produced overt 10 points in the AB.  You can see that the red line races to the top immediately after the opening, crossing the blue line in the process.  That crossing of the blue line is a long entry.  Then the red line stays generally above the 70% threshold and the price continues to rise.   Note how the red and blue lines are separated through most of the up move.  That indicates a strong trend in progress and I would buy more contracts as the price pulls back and hooks.  

When the red line breaks the 50% line you can remove at least one contract if you are trading multiple contracts and if you are trading a single contract, go flat with the crossing of the 50% line.  In the example chart, you’ll notice that the red line again crosses the blue line on its way to the bottom of the window.  You can sell that crossing and go short.  As long as the red line stays below the 50% line you can stay in the short trade.

That’s it in a nut shell for the Aroon indicator.  I do a couple of other things to help with entries and trade management.  Look at the chart below and you can see I have placed the new auto trend lines, recently added to the Ensign program, on the chart example.  I think you can see that it is very beneficial to have these on the chart.  I trade multiple contracts and I use the Trend lines to 'take one off' when the price closes on the other side of a trend line and I reverse the trade when the Aroon red line crosses the blue line, which when it happens correctly, occurs quickly after the trend line break by the price.  Obviously you can reverse the trade if you wish on the trend line break, but I have seen those trend lines do a nasty little adjustment thing on occasion and that’s not fun to get caught in one of those, so I generally wait for the Aroon to confirm the Trend line break and reversal.  It is considered aggressive style to reverse the trend line breaks and you may get whipsawed more often doing that than waiting for the Aroon.

I have included some properties windows for the Aroon. The basic settings are shown there.  As you can see there is a capability to put an “above and below” setting on the red line thus making it turn green when it is above the blue line and red when below it.  That is the second Properties Window shown below.  That makes it easier to manage your trade.  Simply try to be long when it is green and short when it is red.

There will be times when you should stay out of the market.  Those times are when it is chopping.  Look for the price bars to go horizontal and the lines to cycle rapidly with the red and blue lines crossing frequently. You must avoid this type market with the Aroon system.  The Aroon becomes a virtual whipsaw machine in those circumstances.  There are many methods to identify chop.  The action of  the 'Average Bars' is a good way to spot and avoid chop.

When the market is moving normally this system will catch all the winners and cut losers fairly quickly due to the action of the Aroon and the price action around the trend lines.  I use a 0.75 range chart to trade with this indicator.  I’m sure other time frames will work with the Aroon but the settings should be tailored to the time frame.  Give it a try and see if you like it.  A template named the AroonSystem can be downloaded from the Ensign web site.  

I hope this clears up some of the confusion that many seem to have concerning this great indicator.  Good trading to all.

This is the September update of the paper trading model based on Bill Hatch's July article  Straddle-Strangle-Swap.  Last month I defined the scope of this project, and promised to publish an update each month for the rest of 2004.

The model is selling near term options in Dell Computers (DELL), Home Depot (HD), Office Depot (OPD),  and Disney (DIS), and holding longer term options as an insurance policy.   The plan is to execute a calendar roll-out each month by buying back the short options on the day before expiration, and selling short options one month out.   The roll-out of replacing the September options with October options was carried out on Thursday, September 16th, 2004.  The October roll-out will be carried out on Thursday, October 14th.  This is where the model stands.

A desirable characteristic is for the stocks to be tracking sideways.  This was the case in the past month for Dell, Disney, and Office Depot.  The stock that trended was Home Depot, possibly benefiting from increased demand for building materials because of the three hurricanes.

The short HDIG Call at $35 is the trade that was a costly $2000.  The model will resell the same strike in the October month with the expectation that Home Depot prices will return to a more average price.

Summary:

• Dell:    - 125 - 375 + 175 + 775 =     $450 net gain

• HD:     + 975 - 375 - 2025 + 225 = -$1200 net loss

• ODP:  - 525 - 75 + 1175 - 175 =      $400 net gain

• DIS:    - 25 - 325 - 25 + 525 =          $150 net gain