the point I was trying to make, kind of long winded, was that if you back test pretty much any two moving averages as a system i.e. go long when the faster MA crosses above the slow MA and reverse short when the fast MA crosses below the short MA, unless you curve fit them to a given time frame, they will always produce less than 50% winning trades.  Thus logic dictates that yes when MA's cross one another more often than not it's a whipsaw and price will simply resume back in the direction it was already headed.  For example if I test a 13 EMA crossing over a 34 EMA, over one time frame I get only a 33% winning strategy, this means that odds favor you to actually assume the trend will continue in the same direction in most cases when the two MA's cross.  

Best to use much more powerful but simple technical analysis techniques such as patterns, trendlines, symmetry moves, even some rudimentary wave count but only if it's very obvious, and for moving averages they have their place, however best to use them for noting bullish and bearish configuration and price distance from price and each other etc