oh man, do I ever. I was hired once to trade an S&P xover system, this was before auto exec. Everyday we would tweak the parameters, sometimes even 3 times a day. What was working would never work, and what we weren't using started working again.Quote from freealways:
Profitseer, be careful in deciding which set of moving averages.
is the best.
That may be formfitting. Yes it works perfectly on back data but not so hot into the future.
Robustness should be the deciding factor. So one wants to look for sets which are surrounded by nearby sets which all seem to produce good results.
THEN select the one in the middle. Probably not as good a result as a aprticular isolated set perhaps but more robust, more reliable.
I hope you understand what I am trying to say above.
The condtions can be judged by looking for small trading ranges.
Th moment one is in a small trading range one knows there is danger and that it may pay to stay flat.
freealways
Best deal is stick with the same values and trade or fade. And that goes for anything. If you see something fail enough, you sort of get an idea when it IS going to work.
)