Exponential Moving Average The exponential moving average is a more sophisticated moving average. It weighs recent price action more heavily than price action farther in the past but takes into account all the price action in the history of the security with an exponentially diminishing weight. The exponential moving average is calculated a bit differently from conventional moving averages: the difference between the stocks price and the moving average’s level is diminished by a certain factor. In mathematical terms, it is a recurrence relation (each value is defined by preceding values).
The period factor is a constant that influences the speed of the average. It is comparable to the period in a simple moving average. The period factor can be any number between zero and one. The higher it is the faster the moving average reacts, the lower it is the slower the moving average reacts. A way to connect the period factor to the period of a simple moving average is shown below.
For example a 20-day exponential moving average would have a period factor of 0.0952. Now there is a problem with stating the formula since “yesterday’s level” is needed for the calculation. For the first day’s level (which is not defined by the formula) a simple average is taken of the specified period. In other words, the first day’s average is a simple average; from then on it is exponential. From personal research I found that a period factor of 0.13 over monthly plotted S&P 500 closes gives the best return. 
Here a chart showing an exponential moving average of 50 days (period factor: 0.0392) with the respective buy and sell signals at the crossovers. Again, one can see the moving average is powerful in a trending market but near worthless in a sideways moving market like the one that occurs after the last sell signal. Read the article on moving averages to learn about moving average interpretation. |
