As I see it the problem with moving average crossovers(I call them maxo's) is that they look great in retrospect but are trailing indicators in use. Everyone gets suckered in thinking they'll just buy and sell the crossovers. But in real life the price is way past the crossover when it occurs. I suspect many back tested methods using maxo's as buy and sell points that look great in theory are actually flawed due to this. ************************************** "The ability to delude yourself may be an important survival tool." - Jane Wagner
This strategy works but is very slow especially if you use it for exit also. Most of your profit would have been gone by that time.
Of course, which means that you should not be exiting on crossovers, right? So perhaps you should not also be entering on crossovers, right? In other words, that probably does not matter how you enter as long as it is not too much against the trend. That's the basic use of MA crossovers: they tell you the market direction.
Another way to use Ma's is to take a take one when the slope of a Ma changes, i.e enter on the change of the slope of a 15 period Ma, exit on the 8 period ma. Backtesting by Charles Streidsmans finds better results on this method then Ma-crossovers.
You are right, I use something like that too, except that for indicators like TICK and TRIN to anticipate where they can go next and it is not exactly the same, but similar. It's called the inflection, in the scientific parlance.
Yesterday I tried out trading on the slope of a 5 period Dema on a 89tick charts on the NQ. Results were very promessing.
Ponder this grasshopper. If entering on MA crossover and exiting on same MA crossover yields no edge.... then do you really have an edge by using MA's to determine trend?
The slope of a ma is more sensitive than MA-crossovers. i have to admit a combined it with S/R levels.