|
PureBytes Links
Trading Reference Links
|
My initial studies confirm the Osler findings......H & S = 50% accurate.
But does anyone know which TASC issue reviewed the cup-with-handle pattern ?
I have found that pattern can be extremely nebullous to define......which
may indicate why the results were so favorable.
I am attempting to use pattern detection in combination with classic
indicator (%R, RSI, STOCH, etc) to see if the pattern failures can be
predicted.
> -----Original Message-----
> From: Alan Mi [mailto:ami@xxxxxxxxxxxxxx]
> Sent: Monday, December 06, 1999 2:39 PM
> To: omegalist
> Subject: Re: Code for Pattern Recognition: Head and Shoulders
>
>
> tomkohl@xxxxxxxxxxxx wrote:
> > ... Here is a link to some code for pattern recognition of
> a
> > head and shoulders top ...
> > ...
> >> NY Fed have published quite a bit of work looking at the
> >> "head and shoulders" pattern. Carol Osler ...
> >>
> >>http://www.frbsf.org/system/fedinprint/index.html
>
> Osler's research paper shows that this pattern is NOT
> PROFITABLE. That's a very serious pattern recognition
> program she designed and rigorous statistical analysis
> results she derived. Has anyone seen or done this sort of
> pattern recognition work and come up with profitable
> statistical results that contradict hers? Or results for
> other patterns (e.g. bull flags) for that matter? If there
> is a yes answer on a pattern, individual's time spent on the
> quest would be more worthwhile. TASC had a article that
> shows that the cup-with-handle pattern is very profitable,
> among other pattern related articles. It doesn't surprise me
> to see opposite answers to the same pattern due to the huge
> difference in pattern definitions and the levels of
> sophistication of the detection programs.
>
|