Then I think in a wrong direction, okay I will try it once again, anytime.

Meanwhile a more practical question: in the example begin of this thread jcl calculated:

Quote:
vars Trend = series(LowPass(Price,500));

vars Meanness = series(MMI(Price,200));
vars Filter = series(LowPass(Meanness,500));


Is there a relationship between the 500 periods of the LowPass-Filter for the price defining the trend, the 200 of the Meanness and the 500 for smoothing the Meaness? Is there a 'universal' ratio or is it rather a matter of optimization?

Thanks, Sphin