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

Re: Exponential moving average



PureBytes Links

Trading Reference Links




DH wrote:
> 
mcs> >  Since the computation for determining lag is the same for both
simple
mcs> > and exponential averages, why does the exponential average appear
to
mcs> > have less lag than a simple MA? If you chart a 9 period simple,
and a 9
mcs> > period exponential on the same price chart, the exponential very
clearly
mcs> > makes its turns sooner than the simple MA. If this is not showing
that
mcs> > the exponential MA has less lag than the simple MA, what is it
showing?
> 
DH> Mark's example was for a ramp (steadily increasing or decreasing)
input,
DH> not for a turn.



Yes, you're right, he did say, "...when applied to a straight "price"
line..."


DH> If you look at the exponential formula and consider what
DH> happens at a turn you will see the following pseudo code is true
DH> regardless of past price history.
> 
DH> if price < XMA[1] then XMA < XMA[1]
DH> if price > XMA[1] then XMA > XMA[1]
> 
DH> That is not always true for a simple average.
> 
DH> So, if the turn is real, an XMA may catch it sooner. However the XMA
is
DH> more likely to whipsaw based on a single bar price spike.


Yes, this is true too. I appreciate your observations. I suppose at this
point I'll have to search for the methods by which one can ascertain lag
for a given average during a turn. From a practical perspective
(speaking only for myself), the turns are generally of more interest and
value than the steady states.

Regards,
Monte