R.C. Allen developed the 4,9,18sma crossover trading idea a few decades ago. I did a study myself and found using EMAs there were far less 'whipsaw' crossovers that tend to be expensive. When R.C.Allen developed this moving average crossover plan there were no computers in trader's hands of course and everyone used simple moving averages right up into the 80's. With Qcharts I can drag a chart back for years and recently did this with a variety of moving averages comparing simple to exponential moving averages. And the exponential consistently had fewer of these false whipsaw crossover signals. John Murphy also in his most recent work mentioned this and stated when using moving average crossovers for trading signals exponential moving averages were the best. For short term trades he preferred the 5ema-21ema combination. In my own research I soon realized the problem with simple moving averages is they can't react quickly enough the way they are calculated with removing day one and adding today's data. It just takes to long to change the simple moving average line. So what happens is the longer moving average used doesn't move much as it is slow to respond and the shorter moving average not being affected much does move much faster and briefly crosses the longer moving average and then turns back down again producing a false whipsaw. The exponential moving averages don't do this as the longer average used also moves quick enough that the short moving average does not cross over and if it is just a simple dead cat bounce will not cross over for a false signal. I found this with every combination of moving averages imaginable and spent many hours dragging charts backwards and then slowly forward. I used SPY and did them again with higher beta markets and the results were the same. To make the point clear place a 50 and 200 simple moving average on your chart and then a 50 and 200 exponential moving average on as well. Then notice when price reverses the exponential moving averages will respond with price but still track the simple moving averages overall. But the simple 50 and 200 moving averages will just keep right on running in the same old direction for many 'weeks' while price has reversed weeks ago. That is a good acid test for the problem with simple moving averages. The best combination I found was the 10ema/20ema. But a few were close including John Murphy's favourite the 5ema/20ema. One can't ingore whipsaws when using moving average crossovers for trading as they are worse than they look on a chart and are costly.