|
PureBytes Links
Trading Reference Links
|
*Bar-based* Relative Strength Comparative (RSC) vs *date-based* RSC.
Let's have a look at a basic version of yearly RSC between a stock and its
related index.
Assuming today is Fri 23rd Feb 2007, and a year back was Thurs 23rd Feb
2006, the approximate number of daily lookback bars back to that date is
around 252.
Although non-normalized RoC has its problems, we can first take the
approximate yearly Rate of Change (RoC%) of the stock:
StockRoC:=ROC(C,252,%);
Now, we take the same approximate yearly RoC% of its related index:
IndexRoC:=ROC(Security("$SP",C),252,%);
Now, compare their mutual Relative Strength %:
RSC:=StockRoC-IndexRoC;
The problem here is that apart from correlation issues (the stock/index
comparison may or may not be valid), 252 daily bars ago is unlikely to be
the same yearly starting date for both RoCs. 252 bars ago on a daily
chart may be more like a year and a week for some stocks, and much longer
back in time for other irregular stocks.
The StockRoC-IndexRoC Relative Strength Comparison in this (and almost
every) example is likely to be incorrect. ERSA/FIRE uses a related
version of bar-based RSC, and in most cases there will be a mismatch
between common starting dates for its RSC computation. And there lays the
problem with ERSA and "dirty data".
Using a calendar-based starting point (as used in the URSC kit) coupled
with valid correlation (upcoming URSC v4.0), guarantees that the RS
comparison between the two plots will be valid, or at least a very close
approximation where there is no data available for their common starting
date a year back.
Furthermore, because there are no intraday/daily/weekly/monthly variable
bar issues involved, any chart periodicity will produce the same RSC
results with the URSC kit.
There are other potential problems with ERSA-type ranking, such as the
fact that the data folder may have an inherent narrow or wide RSC spread,
yet the ERSA plugin is unable to distinguish between the two. A 90%
ranking based on a narrow difference between securities is not the same as
a 90% ranking based on a broader difference between securities.
Lastly, betting on highly ranked stocks/derivatives almost guarantees a
poor entry point value. The idea of value trading is to catch a strong
stock on its way up, and not when it finally gets there.
jose '-)
http://www.metastocktools.com/URSC/URSC.htm
--- In equismetastock@xxxxxxxxxxxxxxx, superfragalist <no_reply@xxx>
wrote:
>
> If a history file has missing, corrupted or bad data in it, any
> indicator or formula value that is calculated during the lookback
> period that encompasses the bad data--is wrong.
>
> Any signal that an indicator gives during the lookback period with the
> bad data in it is suspect and very likely an incorrect signal.
>
> Maybe there are some new compensation techniques that I don't know
> about that are so good at estimating missing data, having correct data
> is no longer necessary for indicator values to be correctly calculated.
>
> There are many ways to force a formula to make a calculation despite
> any missing, corrupted or wrong data; however, the calculated value
> based on the wrong data is not going to be the same as the value made
> with all of the correct data points.
>
>
>
>
> --- In equismetastock@xxxxxxxxxxxxxxx, "Jose Silva" <josesilva22@>
> wrote:
>
> "Your data has to be clean. If you've got data that's missing bars or
> if it has other problems, the ERS file won't be created properly."
>
> Unfortunately there is practically no security in the world that has an
> uninterrupted historical data record. Just about every stock needs
> to halt trading at one time or another, and this plays havoc with any
> process that involves synchronizing two or more data sets.
>
> Clean data should not be an issue with any properly-constructed formula.
> This is the main reason why I have gone through so much trouble to get
> around the data synchronization problem with the URSC tool-kit.
> The URSC kit works flawlessly regardless of any missing data or chart
> periodicity.
> Data synchronization is essential for any Relative Strength Comparative
> studies.
>
> http://www.metastocktools.com/URSC/URSC.htm
>
> Stand by for the release of URSC v4.0 over the next few weeks, a major
> update with additional custom DLLs and strategies, which will be
> available free to all genuine URSC kit owners as usual.
>
>
> jose '-)
> http://www.metastocktools.com
>
>
> P.S.:
> Perhaps it's high time for
> superfragalist/valuelinetrader/Jim/John/Bill/William/Stan/Bob/whatever
> to come clean and reveal the main reason why he is (apparently) the
> sole apologizer and source of "support" for the ERSA plugin.
>
>
>
>
>
> --- In equismetastock@xxxxxxxxxxxxxxx, superfragalist <no_reply@>
> wrote:
>
> Your data has to be clean. If you've got data that's missing bars or
> if it has other problems, the ERS file won't be created properly.
>
> I run the ERS calculations on 6000 stocks a night and no problems.
> I make sure that no data is required on the last bar in case the
> downloader skips a stock, which it does sometimes. I also clean out my
> history files of stocks that are no longer trading or have been
> removed from the data list for some other reason.
>
> It's all in the data.
>
>
>
> --- In equismetastock@xxxxxxxxxxxxxxx, "Coastal Enterprises"
> <d.stutzman2005@> wrote:
>
> Marion,
>
>
> I've had similar problems. It seems certain securities work and
> others do not. I've been able to figure out what clues suggest a
> workable security.
>
> I know this is not of help correcting the problem, but I didn't want
> to feel alone with the problem.
>
>
> Doug S
>
>
>
> _____
> From: equismetastock@xxxxxxxxxxxxxxx
> On Behalf Of marion dickey
> Sent: Sunday, February 25, 2007 12:05 PM
> To: equismetastock@xxxxxxxxxxxxxxx
> Subject: [EquisMetaStock Group] FIRE Problems
>
>
> Is anyone having a problem with the FIRE ERSA REPORT to work with the
> explorer? When I run the explorer, I get a 100% rejection of all
> stocks explored?
>
> When opening the rejection button, I get the following; " Error in
> column A: An MSX DLL is reporting an internal error. Referenced
> DLL: "ERSA ExternalRelativeStrength" DLL Message: File Not
> Found:\External Relative Strength\CPY.TXT".
>
> I get this regardless of how many record I load in the "Time Periods
> to load field" or the explorer "load options"? I have also reinstalled
> FIRE three or four times with no relief.
>
> Any help, aid, and assistance is appreciated,
> Marion
------------------------ Yahoo! Groups Sponsor --------------------~-->
Transfer from your equities account.
Receive up to $1,000 from GFT. Click here to learn more.
http://us.click.yahoo.com/aZttyC/X_xQAA/cosFAA/BefplB/TM
--------------------------------------------------------------------~->
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/equismetastock/
<*> Your email settings:
Individual Email | Traditional
<*> To change settings online go to:
http://groups.yahoo.com/group/equismetastock/join
(Yahoo! ID required)
<*> To change settings via email:
mailto:equismetastock-digest@xxxxxxxxxxxxxxx
mailto:equismetastock-fullfeatured@xxxxxxxxxxxxxxx
<*> To unsubscribe from this group, send an email to:
equismetastock-unsubscribe@xxxxxxxxxxxxxxx
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
|