|
PureBytes Links
Trading Reference Links
|
It's not a question of using 10 or 20 or 30 years worth of data. It's a question
of possible curvefitting. For example - you might find that a 72 day moving
average system works great. But if 71 and 73 days both stink - then you're
looking at a simple case of curvefitting - not a good system. When you're looking
at time frames in various systems - you should see a curve of results that goes
from good to better to best to better to good (more or less) - like a 50-60 day MA
is good - a 60 to 71 day MA is better - the 73 day MA is best - the 74-80 day MA
is better - and the 80-90 day MA is good. Robyn
Robert Cummings wrote:
> Bollinger used 10 years for his study at least that's what he said.
>
> Robert
|