[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Adjusting a System



PureBytes Links

Trading Reference Links


COMMENTS IMBEDDED

<< This is another case of using indicators blindly and back testing to make 
them
 fit.  An understanding of what these indicators where meant to do and how 
they
 were intended to work would solve this problem.  Both the moving average and 
MACD are trend following systems and are lagging indicators.  
--------------

Actually MACD is NOT necessarily a lagging indicator. It may at times be a 
leading indicator in terms of its signals, and often the lead is wrong to 
boot. Macd parameter defaults on standard software are the results of the 
work and optimization done by Gerald Appel 15 years ago. Certainly nothing 
wrong with a little re-optimization. 

============================================

They tell what the
 condition was x numbers of bars back. Not what is happening now.  The 
assumption
 is that if the trend changed x number of bars back then we are still in that
 trend.  The problem is, which trend are you in?  Each time frame has a 
different
 trend.  First you have to decide which time frame you want to trade. Second 
you
 have to decide which cycle in that time frame you are trading, and only then 
can
 you adjust the indicators to reflect the correct information. The first clue 
is
 that you only use odd numbers for moving averages and things dependent upon
 averages.  Ira >>
-----------------------------------

If you do decide to re-optimize what you will likely find is that there are 
no magic numbers whehther they be odd or even. The market does not know what 
lengths your moving averages are. Continual re-optimization of systems though 
should not effect the performance of most systems over the long haul, as the 
parameter sets that work need be pretty robust (ie the system should work to 
varying degrees over a broad parameter set) if the system is to work at all 
over the long run. Not that the effort is probably not worthwhile for it owns 
sake, but you will likely not find much that really significantly changes the 
results of what you started with. If you do, then you are likely on to 
something, and should probably continue down that path exhaustively.