|
PureBytes Links
Trading Reference Links
|
Roscoe wrote:
>...in the bigger scheme of things it does not pan out. My testing
>has shown that most "smoothers" that look good usually die in the
>a** likewise. Arguments about the lag or overshoot characteristics
>seem to be purely academic to me - my only question is "what are
>the test results?".
Yes, characteristics are academic. When designing a trading
strategy, however, you also have to decide what characteristics you
want in your underlying tools. I can conceive of a few things where
smoothers with low lag and overshoot would be useful as part of
something else, although not as a primary indicator.
My own tests, for example, of statistical probability bands above
and below an expected price value do indicate that a smoother with
low lag and no overshoot represents "expected" price better than
anything else, and the indicator bands that result would not be
used as an entry/exit signal, but as a "reversion to the mean" bias
indicator.
There aren't useless tools, only useless appliers of those tools.
Many indicators, smoothers, and other functions we have ARE useless
for the traditional purposes published about them. But then someone
comes along and finds an innovative way to apply something in a
useful way, when it seemed useless before. Woodie's forum and CCI
is one example.
>> JMA will produce a misterious and unknown spectrum of phase shifts
Raw prices also produce a mysterious and unknown spectrum of phase
shifts. So what?
>Exactly. A small and potentially dangerous 'black box' IMHO.
I don't use black boxes myself. I reverse engineer them when
possible, so they're no longer black boxes. I *almost* succeeded
with JMA, or at least got close enough that I'm pretty sure I know
how it works. The understanding gained from doing that can reveal
the limitations and potentials of the tool.
-Alex
|