|
PureBytes Links
Trading Reference Links
|
On Saturday, March 24, 2001, 12:15:00 PM, Gary Fritz wrote:
>> You bring up an interesting point, and it sounds as though you are
>> (roughly) equating noise with whipsaws.
GF> Well, I wasn't really, but that was the term Ivo used.
Understood. I, also, was referring to the indicator-price relationship
mentioned in the posts.
>> The 'real' definition of noise
>> seems to be a bit elusive. :-) Could you, perhaps, elaborate on this
>> seeming equivalence between noise and whipsaws?
GF> Noise, in the signal-processing sense, is "jitter" or "wiggles" in
GF> the signal you get out of JMA or whatever. A longer length input
GF> results in more smoothing, which gets rid of those extra wiggles.
Agreed. I think you and I look at it the same way, but I was just
checking to see if there was something there re whipsaws and price
location.
GF> Some people try to trade with moving averages (JMA or whatever) by
GF> going long when they turn up, and going short when they turn down.
GF> In that case, any "wiggles" in the signal are going to cause whipsaw,
GF> because they're going to cause you to reverse & reverse again when
GF> you didn't really want to.
Agreed. If you are just looking at change in slope, you don't really
care where price is. It is quite clear that smoothness is very
important in slope-based systems.
I was just wondering if there were any thoughts relating to the
indicator-near(or cross)-price ideas mentioned in the earlier posts.
In those cases, where you are looking at price location relative to
the indicator, the need for great smoothness is not as evident, and
the role of lag seems less clear-cut, and might even be useful. [I am
not, however, suggesting that this is a better approach. :-)] I was
just trying to get a better handle on 'noise', and when, where and how
it is important.
ztrader
|