|
PureBytes Links
Trading Reference Links
|
You may remember that I am looking at real time prices via another product
on a trial basis. I'm using Weekly and Daily bars in Metastock and then
Hourly and 10-minute bars in the other product.
In staring at these various multi-period bars (each with associated
stochastics) for awhile, my current thinking is that one can also think
about a stochastic value at one period confirming the move of the stochastic
at another period. That is to say, if you want to know the significance of
a 10-minute bar stochastic move, say above 20 or below 80, then, look at the
hourly and daily (even weekly stochastic chart.) It seems to me that the
value of the stochastic as an indicator is enhanced by this multi-period
view. (By the way, this is also interesting to look at from the point of
view of MACD and I'm assuming a host of other indicators.)
You may remember that I posted a note recently about an article by Alexander
Elder on the importance of this multi-period view. It was that article that
got me on this particular track. From what I've seen so far, this might
well be a very important insight and one that Metastock in its current
implementation doesn't really help with. (That is, you can't write an
indicator or exploration or system test that references data from multiple
time periods.)
Which brings me to my question: (All my posts that provide info are really
disguised attempts to learn something from someone else.) Has anyone
written an indicator (e.g., a stochastic or MACD) using daily data that
relatively PRECISELY replicates the values of these indicators at the weekly
level? My belief is that the mere increasing of the number of bars within
the indicator doesn't cut it. (e.g., Take a look at the DJIA daily chart
with a stoch(26,3) and then the same DJIA on a weekly with a stoch(5,3).
They're not the same thing.)
Am I wrong in believing that one can't write a stochastic against daily data
that relatively PRECISELY replicates the look of a stochastic on a weekly?
Steven Buss
Walnut Creek, CA
sbuss@xxxxxxxxxxx
-----Original Message-----
From: Frank B. Gaylord <fbg@xxxxxxxxxxxxxx>
To: metastock-list@xxxxxxxxxxxxx <metastock-list@xxxxxxxxxxxxx>
Date: Monday, December 08, 1997 1:08 AM
Subject: Re: MIRAT Long and a Faltering Stoch: A Question
>At 07:17 AM 12/7/97 +0000, Steven Buss wrote:
>>The S&P 500/100 show the daily Stoch(5) and Stoch(13) both well above 80
and
>>the weekly Stoch(5) also well above 80 (and looking like the daily
>>Stoch(13)).
>>
>>The DowInd, Nasdaq Comp, and Nasdaq 100 stochs of the type noted above are
>>below 80 and rising.
>>
>>I assume that this MIRAT Buy Signal, then, would be based on market action
>>in which the Stoch for the two S&P indices would remain above or close to
80
>>for at least as long as it takes the other indices to catch up. Is this a
>>correct interpretation?
>>
>>Frank, my memory is that you're the one who posted the recent note about a
>>Stoch that doesn't drop being trend indicating (or something like that, I
>>can't find the note). (c.f., latest issue of TASC on the
"Pop-Stochastic").
>>
>>Sometimes a Stoch drops below the 80 signal line and this is an extremely
>>important thing to pay attention to because the drop to below or close to
20
>>will turn out to have significant price implications. But sometimes the
>>drop below 80 is very short-term and the return to a reading above 80
>>signals continuation of the trend. (The article in TASC points out that
the
>>return to a reading above 80 is itself a signal of the continuing trend,
as
>>Frank's(?) note also pointed out.)
>>
>>My question is this: Are there one or more indicator(s) one can use in
>>conjuction with the stoch that provides some significantly reliable
insight
>>into whether a below-80 faltering stoch signals the beginning of a
>>significant correction or merely a breather before the stoch moves to
above
>>80 again?
>>
>>In the VERY FEW cases I've looked at, it seems like a default parm MACD
>>provided some early insight into what might be the actual future of the
>>stoch in the kind of case I've laid out above.
>>
>>Thoughts?
>>
>>Steven Buss
>>Walnut Creek, CA
>>sbuss@xxxxxxxxxxx
>>
> :::::::::::::::::::::::::::::::
>
>Steven -
>
> MIRAT does not use Stochastic in its algorithms. Another Tools for Timing
>program, the user interpreted TIMER PROFESSIONAL, does make available
>Stochastic of Price and its version of an Advance/Decline Line. I've
>uploaded to my web site a page showing this. The URL is:
>http://www.usinternet.com/users/fbg/tpstoc.htm. I like what I see in that
>graph, the two Stochastics rising rapidly, with Price leading both. If
>Price were lagging behind, I would not think as I do, that the Market will
>be going up for at least the next few days. I don't have a crystal ball,
>but I do see Price leading the two Stochastics, which will have to do in
>its place!
>
> A couple of notes on Stochastics in this context; MIRAT uses Signal Lines
>of 25 and 75 and you are using 20 and 80. I only point it out for when
>comparing graphs. Also, on my web site page I am showing the DJIA and not
>the S&P 500, so the Stochastics of Price is that of the S&P 500.
>Stochastics of the Advance/Decline Line is of the NYSE, the same with DJIA
>and the S&P 500.
>
> You ask about confirmations of Stochastics. One I use, already mentioned
>above, is Price. Remembering that Stochastics are Moving averages of the
>relationship of Price to the Range of Prices over a given period of time, I
>look at Stochastics as a representation of how well Price is doing against
>that Range of Prices. If Price and Stochastics keep going up, I see
>Stochastics confirming Price. In up movement, it will be Stochastics
>confirming Price, as Stochastics, by its Moving Average nature, must lag
>behind Price, more volatile that Stochastics. Stochastics can lead price
>if Closes are consistently higher from one day to the next and those new
>highs (for that immediate time period, but not necessarily 52-Week or other
>longer than a few days time period) are small. In other words, if Price is
>moving up slowly and each day brings a new high, then Stochastics can lead
>Price in its rise.
>
> If Stochastics snake along between the Top Signal Line (75, 80, whatever)
>and the top of the graph, then I see that as simple confirmation that the
>upward Price move is strong and will probably continue for the next little
>while.
>
> Williams %R and RSI (Relative Strength Index) are two Indicators I have
>used in corroboration with Stochastics. I've also used Stochastics with
>longer parameters, say 21 days and higher and lower Signal Lines (85 or 90
>and 10 or 15) as more of a confirmation of a trend. I should add that I
>only use Signal Lines for points of reference, not saying that something
>has to be done when one is crossed, concentrating more on the direction,
>rate of change and consistency of Stochastics.
>
> I believe that everyone should have a few Indicators or other tools with
>which they become comfortable in using, Indicators that they come to trust,
>and Stochastics is the one for me! For the record, the others are Moving
>Averages and Linear Regression (Line of Least Squares). As I see more and
>more of Jim's Channel stuff, I am feeling more comfortable with Channels.
>
>-- Frank :-)
>
>High Return on Investment using Technical Analysis
>http://www.usinternet.com/users/fbg/
>Minnesota Long Distance Canoe Racing
>http://www.usinternet.com/users/fbg/mnlong/
>------------------------------------------------------------------------
>
|