Here's Soybean Meal with 50 and 200-day MAVGs.
Clearly the crossover signal is slow especially when using longer periods. The middle line on the chart below shows when the crossover goes from positive to negative. It doesn't work well in faster markets as in 2009.
Instead of looking at the crossovers it's often more useful to look at the rate of change for the crossover i.e. is the crossover turning positive of negative. The change in colors above shows when the rate of change for the crossover is positive or negative. Using 50/200-day MAVGs this doesn't generate too many signals -- essentially only seven directional changes since 2008 here. Of course there are fake-outs and dead markets but not too many. How does it backtest? Go to the full version at futures.transmatch.com and pick your commodity and MAVGs and see for yourself.