|
PureBytes Links
Trading Reference Links
|
We speak for stocks here, quite far from this perfect sinusoidal.
If you only imagine the commissions for t=1 you will quickly forget
all about it !!
But, it is not [yet] the end of the world !!
If we were talking about lenses, then we would have to solve the
convexo-concave or concavo-convex surfaces problem!!
At least, we talk for one kind of curvature here.
The subject is interesting, let us go further if you agree.
For example, from some older study, I have pointed out that when C
penetrates the upper Keltner band, the latter is convex and this is
the reason we can expect some profits.
If the upper Keltner band changes suddenly its convex style, then we
are in trouble and the money may fly away. But, before the decline
begins, the curvature will change. With some delay perhaps [Keltner
uses an MA or an EMA] but it will change. It is exactly the point we
need acceleration principle : the curvature will not change at once,
it will pass slowly to the other side and we should take this
advantage to prepare an exit.
The same principle is behind the [recently posted here] TomDeMark
Sequential system. He asks C>Ref(C,-3) for 9 cascade bars, in other
words he searches for a prolonged convexity.
See the examples and come back with [hopefully] bright ideas !!
Dimitris Tsokakis
--- In amibroker@xxxxxxxxxxxxxxx, "sidleysh" <steves@xxxx> wrote:
> Thanks - this looks interesting, but I have to work through the
code
> before I can comment. Obviously what I am trying to do is to enter
on
> acceleration and exit on decelaration (concavity). You are right in
> that the curve has to be smoothed for this to work well, but then
we
> are starting to lag the market, as in all smoothing.
>
> Will get back to you when I understand your code.
>
>
> --- In amibroker@xxxxxxxxxxxxxxx, "DIMITRIS TSOKAKIS"
<TSOKAKIS@xxxx>
> wrote:
> > The absolute convexity criterion would be
> > freq = 1;
> > y=sin( Cum( freq/10 ) );
> > t=1;
> > Convex=(y-Ref(y,-t))/t>(y-Ref(y,-(t+1)))/(t+1);
> > Plot(y,"",IIf(Convex,colorBrightGreen,colorRed),8);
> > Of course, in real life examples this would not help, even in
> smooth
> > curves.
> > Try the
> >
> > y=MA(StochD(50),50);
> > freq = 1;
> > //y=sin( Cum( freq/10 ) );
> > t=1;
> > Convex=(y-Ref(y,-t))/t>(y-Ref(y,-(t+1)))/(t+1);
> > Plot(y,"",IIf(Convex,colorBrightGreen,colorRed),8);
> >
> > Although y is quite smooth, the convexity does not hold for long.
> > At this stage, we should sacrifice the absolute convexity for
more
> > useful conditions.
> > Try t=5 or t=10 to see what I mean
> >
> > y=MA(StochD(50),50);
> > freq = 1;
> > //y=sin( Cum( freq/10 ) );
> > t=1;// t=5 or t=10 or higher
> > Convex=(y-Ref(y,-t))/t>(y-Ref(y,-(t+1)))/(t+1);
> > Plot(y,"",IIf(Convex,colorBrightGreen,colorRed),8);
> >
> > Dimitris Tsokakis
> > --- In amibroker@xxxxxxxxxxxxxxx, "DIMITRIS TSOKAKIS"
> <TSOKAKIS@xxxx>
> > wrote:
> > > An idea for a convexity criterion is the condition
> > > Convex=y>2*Ref(y,-t)-Ref(y,-2*t);
> > > for a given array y and a time segment t.
> > > Try the example
> > >
> > > // Convexity criterion
> > > freq = 1;
> > > y=sin( Cum( freq/10 ) );
> > > t=5;
> > > Convex=y>2*Ref(y,-t)-Ref(y,-2*t);
> > > Plot(y,"",IIf(Convex,colorBrightGreen,colorRed),8);
> > >
> > > in IB, if it is what you ask.
> > > [The non-convex area (red)is the concavity condition, since the
> > above
> > > condition is boolean.]
> > >
> > > Dimitris Tsokakis
> > > --- In amibroker@xxxxxxxxxxxxxxx, "sidleysh" <steves@xxxx>
wrote:
> > > > DT - the problem with this is that it gives a 2 day
> acceleration
> > > > picture - I am looking to find the accelaration over x days
> > (which
> > > is
> > > > easy to see visually - you simply look for a rising convex
> > > pattern).
> > > >
> > > > Any idea how to generalise this?
> > > >
> > > > --- In amibroker@xxxxxxxxxxxxxxx, "DIMITRIS TSOKAKIS"
> > > <TSOKAKIS@xxxx>
> > > > wrote:
> > > > > The Velocity of Close should be
> > > > > Vc=dC/dt
> > > > > and the acceleration
> > > > > g=dVc/dt
> > > > > The minimum dt for daily data is 1 bar, ie
> > > > > Vc=dC
> > > > > g=dVc
> > > > > In AFL
> > > > > Vc=C-Ref(C,-1);
> > > > > g=Vc-Ref(Vc,-1);
> > > > > Plot(g,"g",1,1);
> > > > >
> > > > > Dimitris Tsokakis
> > > > > --- In amibroker@xxxxxxxxxxxxxxx, "sidleysh" <steves@xxxx>
> > wrote:
> > > > > > I am trying to find a way to find out whether, and how
> much,
> > a
> > > > > chart
> > > > > > has accelerted over the previous period. This is not ROC,
> but
> > > > more
> > > > > > true 'rate of change' in a differential calculus sense.
> > > > > >
> > > > > > IN other words - if diff between close today snd close
> > > yesterday
> > > > >
> > > > > > close yesterday and the day before, then the close prices
> are
> > > > > > ACCELERATING up over the last 3 days.
> > > > > >
> > > > > > Can anyone think up a general way to code this?
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Buy Ink Cartridges or Refill Kits for your HP, Epson, Canon or Lexmark
Printer at MyInks.com. Free s/h on orders $50 or more to the US & Canada.
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/GHeqlB/TM
---------------------------------------------------------------------~->
Send BUG REPORTS to bugs@xxxxxxxxxxxxx
Send SUGGESTIONS to suggest@xxxxxxxxxxxxx
-----------------------------------------
Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
(Web page: http://groups.yahoo.com/group/amiquote/messages/)
--------------------------------------------
Check group FAQ at: http://groups.yahoo.com/group/amibroker/files/groupfaq.html
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
|