I just checked some of the roster TS online qualities with my TS 4 version
and found some very different quality ratings.
What would cause such a wide quality variation in what I thought were
relatively similar programs?  Here is a list of several stocks where I found
those discrepancies.

Thanks for your answers.  Carol Lyle

        TK online             TK desktop
ACH   8.4                        3.2
ALJ     9.5                         too young
BHP   6.3                        too small
BNS   5.8                        too young
CACB  9.5                      3.2
CEO  8.4                        2.1
DNA  10                        3.2
HOG  5.8                        2.6
ITU  6.8                           2.1
LOGI  8.9                        3.2
RIO  8.9                         3.2

Carol, et al:

At 12:49 PM 11/7/2007, you wrote:

I just checked some of the roster TS online qualities with my TS 4 version
and found some very different quality ratings.
What would cause such a wide quality variation in what I thought were
relatively similar programs?  Here is a list of several stocks where I found
those discrepancies....

The issue is this: the source of data from which the original Complete Roster
of Quality Companies is harvested has been OPS/SDS from BI until this issue.
While the same algorithms are used to provide the judgmental decisions that
are made (and parallel those in the Take Stock programs, whether desktop
or on-line) and from which the Quality Indices are constructed, the differences
in the actual data will cause different conclusions to be drawn.

This month, for the first time, the StockCentral (Hemscott) data was used to
create the Complete Roster. And, when you run your Take Stock programs
that still use the OPS data, the results are going to highlight the differences
in the data.

I have nearly completed a thorough study of the various sources of electronic
data and will soon publish it. The result of that study shows that the number
of discrepancies to be found in each source (Reuters, Compustat, and Hemscott)
appear to be fairly evenly divided amongst all three. Just because we happen
to have "grown into" believing that the Compustat data is the "gold standard"
doesn't necessarily mean that it is always right and the other sources always
wrong. Quite the contrary. All are "right" most of the time.

As I will explain in my report, there are a number of levels at which the data
can be in error, beginning with simple typos, which are to be found in all of
the sources from time to time, to differences in the number of years/quarters
of data that are provided (some providers, for example, will restate data
when there is a merger while others may not even offer the data prior to
that happening). Some providers use the leeway provided by the GAAP
standards to modify the data to benefit their clients; i.e., will remove the
non-recurring or extraordinary data to please analysts, while another source
will provide the most painstakingly accurate historical data to please their
accountant clientele. Some of the providers will put unusual data in footnotes
while others will insist on retaining it in the tabulations. Some providers take
pains to dredge the footnotes for the unusual data, and then eliminate it in
their normalizing process, while others don't do either.

All of these results are "right," even though they are different. After the data
leaves the source providers, there are also differences between StockCentral,
BI, and AAII, when it comes to the data they select to populate the individual
line items we require. Again, these represent differences of opinion as to what
data is more appropriate to publish. As an example, there are those who think
that the pre-tax profit should be normalized, while others reserve normalized
results for the Earnings per Share.

Each of these differences that are found at different steps in the process of
producing usable data can cause a difference in the end result. And, because
the Quality Index is a compounded result of all of those differences coming
together at one point, those differences are magnified as the results emanate
from the different sources.

Should the BI data continue to be used for the roster, then the results you
arrive at using the Take Stock versions that use OPS data will match up; and
you will be content. Or, when the Roster derived from StockCentral data is
analyzed using the on-line version of Take Stock, the same result will occur.

It's only when you mix the data from this Roster and the later analysis, that
you find those disconcerting differences. Or, you use StockCentral's screener
to collect all those with an acceptable or desirable Quality Index (which is, in
essence, the same process as the Complete Roster uses) and then use the
desktop versions of Take Stock to analyze them that you find such differences.
And it was just this set of circumstances that motivated me to commence this
study.

I will make every effort to get this completed and available for publication as
soon as possible. I've had a number of things that have drawn my attention
away from this effort in recent weeks; but those have pretty well subsided
now and I can go back to work on it.

For the time being, I would suggest that you try to use only those companies
in the Roster or the StockCentral screens that agree in their Quality Indices
no matter which version of Take Stock you use. That way, no differences in
the data will affect the outcome and you can have double the confidence in
the result.

ET

----- Original Message -----

From: toolshed@stockcentral.com

To: clyle@columbus.rr.com

Sent: Wednesday, November 07, 2007 1:28 PM

Subject: [The Toolshed]: RE: Take Stock online vs. Take Stock desktop (55fc04bd-0824-409b-9ce2-2552553841ca)

From the The Toolshed forum at StockCentral.com, Doug Gerlach writes:

The discrepancies come from the different data sources, as well as the dates of the most recent final data, that are used in the programs. At some point, we'll possibly change the ICLUB Take Stock to use the StockCentral data instead of S&P, and that ought to bring the results closer together.

Doug

----------
Posted by: Doug Gerlach

----------
To view the complete topic, reply, or unsubscribe to this topic please visit: http://www.stockcentral.com/community/tabid/143/view/topic/postid/4288/ptarget/4289/language/en-US/Default.aspx

---

No virus found in this incoming message.
Checked by AVG Free Edition.
Version: 7.5.503 / Virus Database: 269.15.24/1115 - Release Date: 11/7/2007 9:21 AM

----- Original Message -----

From: toolshed@stockcentral.com

To: clyle@columbus.rr.com

Sent: Wednesday, November 07, 2007 2:39 PM

Subject: [The Toolshed]: RE: Take Stock online vs. Take Stock desktop (55fc04bd-0824-409b-9ce2-2552553841ca)

From the The Toolshed forum at StockCentral.com, Ellis Traub writes:

Carol, et al:

At 12:49 PM 11/7/2007, you wrote:

I just checked some of the roster TS online qualities with my TS 4 version
and found some very different quality ratings.
What would cause such a wide quality variation in what I thought were
relatively similar programs?  Here is a list of several stocks where I found
those discrepancies....

The issue is this: the source of data from which the original Complete Roster
of Quality Companies is harvested has been OPS/SDS from BI�until this issue.
While the same algorithms are used to provide the judgmental decisions that
are made (and parallel those in the Take Stock programs, whether desktop
or on-line) and from which the Quality Indices are constructed, the differences
in the actual data will cause different conclusions to be drawn.

This month, for the first time, the StockCentral (Hemscott) data was used to
create the Complete Roster. And, when you run your Take Stock programs
that still use the OPS data, the results are going to highlight the differences
in the data.

I have nearly completed a thorough study of the various sources of electronic
data and will soon publish it. The result of that study shows that the number
of discrepancies to be found in each source (Reuters, Compustat, and Hemscott)
appear to be fairly evenly divided amongst all three. Just because we happen
to have "grown into" believing that the Compustat data is the "gold standard"
doesn't necessarily mean that it is always right and the other sources always
wrong. Quite the contrary. All are "right" most of the time.

As I will explain in my report, there are a number of levels at which the data
can be in error, beginning with simple typos, which are to be found in all of
the sources from time to time, to differences in the number of years/quarters
of data that are provided (some providers, for example, will restate data
when there is a merger while others may not even offer the data prior to
that happening). Some providers use the leeway provided by the GAAP
standards to modify the data to benefit their clients; i.e., will remove the
non-recurring or extraordinary data to please analysts, while another source
will provide the most painstakingly accurate historical data to please their
accountant clientele. Some of the providers will put unusual data in footnotes
while others will insist on retaining it in the tabulations. Some providers take
pains to dredge the footnotes for the unusual data, and then eliminate it in
their normalizing process, while others don't do either.

All of these results are "right," even though they are different. After the data
leaves the source providers, there are also differences between StockCentral,
BI, and AAII, when it comes to the data they select to populate the individual
line items we require. Again, these represent differences of opinion as to what
data is more appropriate to publish. As an example, there are those who think
that the pre-tax profit should be normalized, while others reserve normalized
results for the Earnings per Share.

Each of these differences that are found at different steps in the process of
producing usable data can cause a difference in the end result. And, because
the Quality Index is a compounded result of all of those differences coming
together at one point, those differences are magnified as the results emanate
from the different sources.

Should the BI data continue to be used for the roster, then the results you
arrive at using the Take Stock versions that use OPS data will match up; and
you will be content. Or, when the Roster derived from StockCentral data is
analyzed using the on-line version of Take Stock, the same result will occur.

It's only when you mix the data from this Roster and the later analysis, that
you find those disconcerting differences. Or, you use StockCentral's screener
to collect all those with an acceptable or desirable Quality Index (which is, in
essence, the same process as the Complete Roster uses) and then use the
desktop versions of Take Stock to analyze them that you find such differences.
And it was just this set of circumstances that motivated me to commence this
study.

I will make every effort to get this completed and available for publication as
soon as possible. I've had a number of things that have drawn my attention
away from this effort in recent weeks; but those have pretty well subsided
now and I can go back to work on it.

For the time being, I would suggest that you try to use only those companies
in the Roster or the StockCentral screens that agree in their Quality Indices
no matter which version of Take Stock you use. That way, no differences in
the data will affect the outcome and you can have double the confidence in
the result.

ET

----------
Posted by: Ellis Traub

----------
To view the complete topic, reply, or unsubscribe to this topic please visit: http://www.stockcentral.com/community/tabid/143/view/topic/postid/4288/ptarget/4291/language/en-US/Default.aspx

---

No virus found in this incoming message.
Checked by AVG Free Edition.
Version: 7.5.503 / Virus Database: 269.15.24/1115 - Release Date: 11/7/2007 9:21 AM