|
PureBytes Links
Trading Reference Links
|
Hallo Fred, Lionel and Jan Willem
I didn't miss out on anything, but you are right as to what the article was
all about. Not only Williams %R(wasn't that presented "upside down" in MS6.0 ?)
but also more momentum etc. indicators have hit the good and extreme OB/OS
levels when Price was/is reaching its accelerating or turnaround "points".
I have used a simular system as McKinnans, and found it to be usefull to futher
look into, but have also have had times when it didn't predicted right, and as such
costing a bit of money too.
Also you are right that they are too wide, but also, here comes: Never Hitted.
The bottom Line does occaisonally at a for that period "all time low" but I presume
does so by accident.
As such I wasn't revering to your mails basic content or McKinnans good article, but
to the other Raff RC Tool being used as articles base.
Raff RCs
What does struck me again and again is to find users(also among our group
overhere) that are even still using the Raff floating channels for their trade dissisions.
These users have commented, and in this I agree, that these build-in Channeling Tools
were so easy to use and "helpful" when used. But when they present false information,
then its no wonder they find it all too complicated and move away from the markets.
If you haven't got the right tools, the only "possible" gessings remains.
Metastock refers in their build in Help files and manuals that the LRC lines and
the Linear Regression Channels do come standard build in with the software.
In this they refer to the following Tools:
A.-the Linear Regression as Indicator
B.-the Linear Regression as Line Tool
C.-the Raff Regression as Linear Regression Channels Tool
>From the Helpfile/manual
-quote-
Linear Regression Trendline
-A Linear Regression trendline uses the least squares method to plot a straight
line through prices so as to minimize the distances between the prices and the
resulting trendline.
-A Linear Regression trendline is simply a trendline drawn between two points
using the least squares fit method. The trendline is displayed in the exact middle
of the prices. If you think of this trendline as the “equilibrium" price, any move
above or below the trendline indicates overzealous buyers or sellers.
-A Linear Regression trendline shows where equilibrium exists. Raff Regression
Channels (see Raff Regression Channel) show the range prices can be expected
to deviate from a Linear Regression trendline.
-Linear Regression trendlines is used to construct Raff Regression Channels
(see Raff Regression Channels), Projection Bands (see Projection Bands),
Projection Oscillator (see Projection Oscillator) and the Linear Regression indicator
(see Linear Regression Indicator).
Raff Regression Channel
-The Raff Regression Channel is constructed by plotting two parallel, equidistant lines
above and below a Linear Regression trendline. The distance between the channel
lines to the regression line is the greatest distance that any one high or low price is
from the regression line.
-unquote-
Now note in this that the Raff floaters are purely based on Equis version of how to draw
and then calculate a Linear Regression Line.
-1. From missing the right Start point(see futher below) this explains "Part-1 of 2" of their
diverging or drifting away from a standard Linear Regression Line.
-2. From the "greatest distance that any one high or low price is from the regression line"
and when plotted the Raff floaters do not ever at one side holding on to this Greatest
recorded distance, "Part-2 of 2" is explained, and also I can therefore asume that they
are daily(per each day) recalculated, and as such dynamic. This dynamic does not
stand for powerfull, but for moving!!!!!!!!(=floating, especialy in this: they start drifting).
Linear Regression Channels over time are not dynamic, but static, using the extremest
recorded highest High or highest Low and/or up to users preference the highest and
lowest Close, as their outer borders, as that is where the price turnarounds took place.
If new, greater Highs or Lows are made by the Price then a readjustment of the Start point
is to be made and/or this could involve to a move up in time or in an up or down level
readjustment to be made. Otherwise, daily set highs or lows(or higher or lower closes) are
irrelevant, eg of no futher statistical relevant value to LRCs.
Futher comments:
A.The Indicator does its job allright. Its daily calculated on that days values and as such
dynamic(=oscilate).
B. But the Regression Line is an Equis Regression Line. If you place the indicator at
one starting point somewhere on the chart, lets say a major Low, then it should stick to
this Low point, NOT CALCULATING THAT DAYS VALUE and only for the NEXT and following
days start with counting the Linear values and then to show the in the future established pricings
within a 'middle' Linear Regression Line.
For starters the Equis version drift upwards on the first starting day, as you drag the study to the
right, entirely lifting up and missing out on users original base point, the Start POINT.
(Note in this and as such that the build-in Stdev and Stderr Channels do the same lifting and as
such screw up too!)
This statistical Start point doesn't nescaceraly have to be the Mid Price of that day, neither a
High, Low or Close or even that days Exchange recorded Price. It is just a point somewhere
on the chart.
Then the very important development starts AFTER that day and in time in the future.
So that days point somewher in the chart is only a statiscal point to start from. Now look as you
drag the study into the future, that it uses multi 'start' points within that day. Non equal as to where
you, the user, wanted it to start.
Also this Start point should always, especialy for the users sake, be given with a choise as to
the way and if user would like this to be calculated on a daily, weekly, monthly etc. basis and also
there should always be a choice given to user as to the way and if user would like to use the
extremes -ever recorded- for the calculation, eg the highest High and lowest Low for that period
only when dragging the study, and thus filtering out the glutter and noise of the daily established
and recorded unvaluable higest High and lowest Low, or the use another way like the build-in
Equis version does: calculating the everyday daily established Highs and Lows.
Extreme ever recorded highest High and lowest Low(one of each in a period used)
Besides for historical information, for any further future projection, eg stretching or extending
the Lines to the right when in the chart, the extreme ever recorded highest High and lowest Low
one of each in a period used are as a source in this instance of very vital importance to its
leveling out use for in the near future, also as its used elswhere in other measuring tools,
usabilities and usages, and of very valuable vital information for the users reading or futher
channeling on charts or in creating trading systems.
All Raff floating lines were diverging as is shown before in the previously attached charts.
Drawing both a standard and the Equis version shows Equis version to be a Linear Divergence
Regression. Perhaps usefull as an idea for an entirely new indicator to be 'build in', but then again
this build-in was already partialy done so, but not as such naming the part.
C. The Linear Regression Channels are not at all included with the program software.
Whoever Mr Raff is, is not of real concern, but like as it is with the build-in or any custom indicators,
its either his version or Equis version, but not ever the standard Linear Regression Channels.
As such the same critera as for creating a Linear Regression Line goes up for creating
any of the Linear Regression channels. Parameters such as Base=Base and High=High
and Low=Low as the users choises, wich are everywhere statisticaly and technicaly used, are
missing. And thus too, the variaty in periodical periods to be configured by user.
Regards,
Ton Maas
Ms-IRB@xxxxxxxxx
-----Oorspronkelijk bericht-----
Van: Fred Bender <fbender@xxxxxxxxxxx>
Aan: metastock@xxxxxxxxxxxxx <metastock@xxxxxxxxxxxxx>
Datum: dinsdag 27 oktober 1998 4:01
Onderwerp: Re: Raff channels
>
>
>A.J. Maas wrote:
>
>> As said the Raff Regression Channels are daily RECALCULATED AND therefore
>> CHANGING DAILY. Trends do not excist per day but in an overal period. In this overal
>> period extremes are hit upon, registered and when giving too much resistance will
>> lead into the price bouncing off. An extreme is set. This extreme point will be the
>> Linears basic value for the future. Daily highs and lows, open and closes are NOT
>> these extremes, but only a part of the long road wich will lead up reaching these
>> extreme points, the secundairy movements. The primary move/values in/over time
>> is Linear Regressions basics.
>> ------------------------------------------------------------------------
>>
>> Name: Floating.gif
>> Floating.gif Type: GIF Image (image/gif)
>> Encoding: base64
>
> Ton,
>
>I think you're mssing something. The technique to which I referred, from the Nov.98 TASC,
>is to define the Raff channel based on two points and then extend it out toward the right
>(future time, prices). The author's claim is that frequently the next rally or decline will
>bounce off the lines projected into the future. These lines will not bounce around on a
>daily basis once they've been fixed.
>
>I make no claims for the technique working. I was just struck by how many such llines are
>converging only a handful of points above the presentlevel of the S&P 500.
>
>Regards,
>Fred Bender
>
>
|