# McGinley Dynamic and Moving Averages

McGinley Dynamic
The McGinley Dynamic is a market tool invented and perfected after many years by a 40 year trader, a 40 year market technician, who can add to his lengthy credits, a long time officer of the Market Technicians Association and former Editor of the their Journal of Technical Analysis.
This article will introduce traders to his little known tool first published in the Journal of Technical Analysis as an outline in 1991 and later published as a full blown study in 1997. My hope is to capture the mind of a master technician as he worked through the process of invention to perfection of the McGinley Dynamic who has advanced the study of technical analysis more than most would know. We begin with moving averages.
Moving averages is the study of the history of  time series analysis. Early practicioners used various algorithms to smooth data and to flatten varied shaped curves. Yet this early work was quite primitive. Various graduation methods were used later like a fitting line using a least squares rule for plotting and construction purposes. Fitting lines using the least squares rule was later adopted in technical analysis in the family of moving averages.This began the process of interpolating data using probability theories and analysis.
In the Journal of the Royal Statistical Society in 1909, G.U. Yule described instantaneous averages that R.H. Hooker calculated in 1901 as moving averages. Yule identified properties of the variate difference correlation method.  The term moving averages entered the lexicon it was said shortly after in 1912 through W.I King’s  publication of Elements of Statistical Method.
Harold Wold later adopted Yule’s studies and described moving averages in a 1938 study called the Analysis of Time Series.
Others attribute exponential smoothing to Brown and Holt late 1950’s book about inventory control. Brown used exponential smoothing for Naval inventory processes. Holt was the first to use linear and seasonal trends for inventory control.
Pete Haurlan was the first to use exponential smoothing for tracking stock prices and advanced the study for technicians in the modern day. Haurlan called exponential smoothing trend values where a 19 day EMA he called a 10 percent trend. His earlier work as a designer of tracking systems for rockets helped him design steering mechanisms. If the steering mechanism was off, it needed further inputs. Haurlan called this proportional control and used this method in his groundbreaking studies.
For Haurlan and others, EMA’s was the moving average method of choice because of its focus on two inputs as opposed to the simple moving average that needed many past data points. These early technicians used pure mathematical calculations graphed on chart paper.
For example,  Haurlan needed a conversion factor, a smoothing constant. His smoothing constant = 2/ (n+1) where N is the number of days.
So a 19 day EMA equates to a 10 percent trend  by 2/ (19+1) = 2/20=0.10 or a 10 percent smoothing constant. Proportionate control  equates to how far price moved from the trend value and adjusts by using trend value curves. This he charted in waves by 1%, 2%, 5%, 10%, 20%, and 50 % .
Haurlan developed tracking rates based on trends. These tracking rates were measured against a stabilization period. For example, a 50 % tracking rate has a 5 day stabilization period.
Sherman and Marian McClellan added two different EMA’S of daily breadth figures, 10 % and 5 % trend. This gave the first alerts to crossovers when the 10 % trend moved above the 5% trend. This detected a market reversal as well as overbought and oversold markets.
The McClellans would later invent the McClellan Oscillator and the Summation Index based on their calculations and charting methods during this period and published in their 1970 book, Patterns for Profit. The McClellan Oscillator measures the acceleration of daily advance decline statistics by smoothing with two different EMA’s and finding the difference between the two.
For Haurlan and Loyd Humphries after him with his groundbreaking book called the Moving Balance System and his invention of the Moving Balance Indicator, both benefited from the easier use of coding EMA’s that only needed two inputs, price, angle and position and the prior value.
Back then, computer sophistication wasn’t available. Hence the reason for EMA’s over Simple Moving Averages that needed many data points. What separated McGinley from earlier technicians was his groundbreaking work in moving averages, following where others left off, that led to the McGinley Dynamic.What did he see. I paraphrase.
McGinley says the problem with moving averages is twofold, inappropriately applied and overused.
They should only be used as smoothing mechanisms rather than a trading system and signal generator. Consider as he said, moving averages range in their uses from fast to slow markets. How can one know which to use and appropriately apply them. How can one know when to use a 10 day average from a 100 day. Further, moving averages are fixed in length without ability to change, a restriction in its use because it can’t adjust to changing data during trading days. We know lengths today as slopes.
The hope is the ability of a smoother to filter whipsaws but outliers exist in the averages. What should you do with a 10 day moving average on the 9th or 10th day. It doesn’t work because much of the trend has been lost.
Next says McGinley, simple moving averages are always out of date. A 10 day average is off by 5 days or half its length and graphed wrong. Chances are big price moves already occurred within the 5 days so the graph set at 10 periods must also be off.
The further problem is the drop off, the difference in price and the line. What if the new data from X days ago is dropped and the data drop is larger than present values. The moving average must also drop generating false signals.
Next, exponential moving averages where much is directly quoted so I can replicate modern day examples.
The exponential moving average improves on the simple moving average because calculations allow the average to hug prices more smoothly and allows for faster response to market data. Yet it under performs in consolidations just as the simple moving average generating line breaks and sheer trading indecision.
Exponentials require two inputs, previous average and current price. The classic calculation is A X the previous moving average + B X new data where A+B = 1.0.
Usually a small part of the new data is added to a large piece of the old. To build on the earlier works of Haurlan, for example, an 18% exponential  where A=0.82 and B=0.18 can be compared to a normal moving average where B=2/(x+1).
So an 18% exponential (x=10) hugs prices as closely as a 10 day moving average (2/(10+1)= 18. The shape of the exponential may be different due to calculations. The exponential calculation of B can be adjusted to fit market data and prices where simple moving averages are fixed, its much more rigid due to its calculations.
Exponential moving averages therefore follows prices and market changes better than a fixed simple moving average, it smooths the data better. Yet the exponential moving average is not perfect, adjustments are always needed and it can’t rise with falling prices and fall with rising prices. So what’s the answer, enter the McGinley Dynamic.
Building on years of moving average research, the McGinley Dynamic was invented as a market  tool designed to generate less whipsaws, hugs prices more closely, adjustable calculations to fit the users needs and follows markets fast and slow automatically.
Think of the Dynamic Line of the McGinley Dynamic as Haurlan’s steering mechanism, a proportionate control tool that steers the Dynamic Line along with prices. The questions whether the McGinley Dynamic lives up to its reputation, the answer is unquestioningly yes. Does it perform the above functions, absolutely. Here’s how. Again I quote.
Building on Dr. Lloyd Humphrey’s work of moving averages in his groundbreaking 1976 book The Moving Balance System where the previous Dynamic Line was modified, here is the new formula.
New Dynamic= Dynamic +(index-Dynamic -1)/(N X(Index/Dynamic -1) 5. The index may be the Dow, S&P or a stock.
Mr. McGinley divides the difference between the Dynamic and the index by N times the ratio of the two. The numerator gives the up or down sign and the denominator stays within percentages within the bounds defined by N.
McGinley further states the 4th power gives the calculation an adjustment factor that increases more sharply the greater the difference between the Dynamic Line and the current data. Quoting further, the size of the adjustment changes not linearly but logarithmically. This feature allows the Dynamic to hug prices.
Mr. McGinley recommends N should be 60 % of the moving average one wishes to emulate. His example is a 20 day moving average that uses an N of 12. Herein the Dynamic Line adjusts itself by speeding up or slowing down as markets dictate. The second term of the equation McGinley states is not a factor unless the difference between the index and the Dynamic Line is large. This aspect of the equation deals with lengths or slopes.
An important factor is the second term however. McGinley says in fast up markets, the Dynamic Line slows  down less than down markets. Its the factor of the 4th power that speeds up the Dynamic Line in down markets.
From McGinley’s example, insert 10 for the old Dynamic, 5 for the close and N=7, a product of -6.67. Further, make the close =14 and you get 0.15.
So 14 is far above the old Dynamic as 6 is below. Not a problem says McGinley as the object is to let profits run and bail out when the market drops. So upside profits run without whipsaws while the downside adjusts quickly to a drop allowing opportunity to cut losses. So do you avoid a loss or grab a gain must be the question here and decision for intended users of the Dynamic.
Exactly what does the McGinley Dynamic do.
Mr McGinley set out to avoid whipsaws as moving averages are prone to do and find a tool that won’t separate prices from the average.
In this instance, it avoids large drop offs. The Dynamic Line rises with falling data. Only one piece of back data is needed.
In any trending or trading market, the Dynamic doesn’t need back testing or adjustments. In instances of extreme whipsaw markets, it still sells high and buys low. The main point is moving averages get separated from prices. What happens when a crossover occurs. One may have a loss. So the Dynamic avoids these dilemmas.
Notice the term market tool used throughout this article. Mr. McGinley says the Dynamic Line is not an indicator and shouldn’t be used as such.
He abhors the idea using the Dynamic as a trading vehicle. This was his purpose to comment regarding the problems of the ratios of the up and down markets. Rather I believe, maybe the intention is it should be a market tool to gauge where the market may be in relation to other market tools used by traders. Just speculation.
It should also be noted, the McGinley Dynamic is not only a remarkable tool but its the product of many years of intense research and insight by a master technician.
The author would like to thank John McGinley, a good man, for decency, patience and understanding to allow time to get it all right to bring forth the McGinley Dynamic.
The author would like to offer many thank you’s and a debt of gratitude to Tom McClellan Editor of the McClellan Market Report and McClellan Financial Publications for access to loads of research.
Suggested reading:  Colby and Meyers Encyclopedia of Technical Market Indicators, 1988.
Dr. Lloyd Humphrey, The Moving Balance System, Windsor 1976
Pete Haurlan Measuring Trend Values 1968– Details can be read at Mcoscillator.com.
Sherman and Marian McClellan, Patterns for Profit 1970– Details can be read at mcoscillator.com
November 2009 Brian Twomey

Brian Twomey is a currency trader and adjunct professor of Political Science at Gardner-Webb University