Login
Welcome Guest! To enable all features please Login. New Registrations are disabled.

Notification

Icon
Error

Options
Go to last post Go to first unread
heberling  
#1 Posted : Friday, November 7, 2008 3:12:17 AM(UTC)
heberling

Rank: Newbie

Posts: 3

In a few hours of evaluating the TK6 demo, I have found the new features very useful, and I am still discovering them.
 
However, the results in TK6 seem to have large errors in two sections. In section 1, the trend lines don't appear correct.  As one example of the problem, in analyzing Gilead Sciences (GILD), if I delete the years 1998 through 2002 as outliers, the resulting trend lines don't appear to fit the remaining data at all. The pretax profit and earnings per share trend lines are especially bad. The resulting historical sales growth rate is 63.9%. The resulting historical earnings per share growth rate is 180.8%. If the same data is analyzed using TK5, the trend lines do appear correct and the resulting historical sales growth rate is 49.1% and the historical earnings per share growth rate is 47.7%.
 

The second problem is in section 2B. The TK6 GILD data for % Earned on Equity are very different from the same data in TK5.  There is a note in the TK6 row header that it is previous year data but that does not explain that large difference.  The TK5 data appears to be consistent with Value Line.

I could probably attach images of the results, but it will only take a few seconds to duplicate my example.

Wanna join the discussion?! Login to your forum account. New Registrations are disabled.

GeneBR  
#2 Posted : Friday, November 7, 2008 6:26:05 AM(UTC)
GeneBR

Rank: Advanced Member

Posts: 62

Was thanked: 1 time(s) in 1 post(s)






Paul, I had previously reported the same thing, on
FLIR I believe.  It could really throw your calculations off if you should
choose end of trend line for your starting place for projections.  It isn't
on all stocks when you choose outliers, but I have seen it off and on on some
others, as well.  I hope they are already working on it, but if you are
not, guys, please get on it.    Gene Rooks


elrekce  
#3 Posted : Friday, November 7, 2008 4:33:42 PM(UTC)
elrekce

Rank: Advanced Member

Posts: 34

Paul, 
I agree with your find along with Gene Rooks that the TK6 section 1 trend lines are askew from what they are in TK5 and also believe there is a bug here in TK6. We see this with BI's S&P Data but it may be in the SC M* data also.   I am hoping  that there will be some sort of acknowledgement soon that it is recogonized as a bug and being worked on or an explanation for validity of the difference between the two Toolkits. 
 
In reference to your second point, (I should have recognoized this at our meeting last night) the data in TK6 is programed to report prvious year data as footnoted.  In TK5 you have the option of showing the same previous year data with the hidden keys "Alt + C".  When this is selected in TK5 the 2b row is identical to TK6.  Ellis Traub has written an article concerning the logic for the different data reporting.  If someone can point a reference to that article, it will give a better explanation than I can.

 

Marty Eckerle

gerlach  
#4 Posted : Monday, November 10, 2008 7:36:50 AM(UTC)
gerlach

Rank: Administration

Posts: 437

Thanks: 7 times
Was thanked: 14 time(s) in 10 post(s)
Posted By Paul Heberling on 11/07/2008 8:12 AM

In section 1, the trend lines don't appear correct.  As one example of the problem, in analyzing Gilead Sciences (GILD), if I delete the years 1998 through 2002 as outliers, the resulting trend lines don't appear to fit the remaining data at all. The pretax profit and earnings per share trend lines are especially bad. The resulting historical sales growth rate is 63.9%. The resulting historical earnings per share growth rate is 180.8%. If the same data is analyzed using TK5, the trend lines do appear correct and the resulting historical sales growth rate is 49.1% and the historical earnings per share growth rate is 47.7%.

I wasn't able to replicate that at all. I opened GILD and eliminated the outliers for 1998 through 2002 and the trend lines were fine, with the same growth results that you cited in TK5.  What version of TK6 are you using? It may be that this issue was fixed in one of the updates we've released.

The second problem is in section 2B. The TK6 GILD data for % Earned on Equity are very different from the same data in TK5.  There is a note in the TK6 row header that it is previous year data but that does not explain that large difference.  The TK5 data appears to be consistent with Value Line.

This is by design -- if you have the Preference set to use Beginning of Year Equity in the Section 2b calculation, then the calculations are made using the previous year's End of Year Equity (which is the next year's Beginning of Year Equity). The concept is that if you're measuring the RETURN on Equity, then you should compare the end of year EPS to the company's equity at the beginning of the year. This is an improvement over the way that the ROE calculations have been done in the past, but if you don't like it, then you can turn it off in the Preferences.

Doug

heberling  
#5 Posted : Monday, November 10, 2008 8:46:18 AM(UTC)
heberling

Rank: Newbie

Posts: 3

Doug,

Thanks for looking into this.  I am new to TK5 and TK6 so I may be doing something wrong.  But I have seen the problems in TK6's sections 1 and 2B on another computer with an experienced user running the program.  My TK6 is version 6.0.2.  It was downloaded only last week. I will attach two screen grabs of the TK5 and TK6 results side-by-side in the hope you can see something, especially when you compare my results with the results on your computer.

Paul Heberling

GeneBR  
#6 Posted : Monday, November 10, 2008 3:14:31 PM(UTC)
GeneBR

Rank: Advanced Member

Posts: 62

Was thanked: 1 time(s) in 1 post(s)






Paul, I wrote earlier that I had seen your problem
of unexpected results AFTER eliminating outlier years.  In particular FLIR,
which still shows the same aberration, that Irving has already looked at and called me about.
 
But as Doug reports, the GILD example is
giving you the cockeyed results you reported before, not after, eliminating
outliers, at least in my program.  Having said that, I'd have to keep
my eyes open for  other examples, but I have seen unexpected trend
lines on occasion with other companies as well, after eliminating outliers. 
 
It is good to keep your eyes open for these things to avoid mistaken judgments.  Gene Rooks


gerlach  
#7 Posted : Tuesday, November 11, 2008 4:24:55 AM(UTC)
gerlach

Rank: Administration

Posts: 437

Thanks: 7 times
Was thanked: 14 time(s) in 10 post(s)

I'll repeat what I said earlier about Section 2B -- the data you see in TK6 is enabled by default and is intended.

In TK5, you can go to page 2 and press Ctrl-Alt-R to toggle the option on and off to use Beginning of Year or End of Year equity in the calculations there. If you do, you'll see that it matches TK6 exactly. If you don't like this, you can turn it off in TK6 Preferences > Stock Study > Use earlier year instead of current year in 2B.

The attached excerpt from Ellis Traub's Take Stock book explains some of the rationale behind this.

I will continue to investigate the trend line issue.

Doug

 

gerlach  
#8 Posted : Tuesday, November 11, 2008 4:29:13 AM(UTC)
gerlach

Rank: Administration

Posts: 437

Thanks: 7 times
Was thanked: 14 time(s) in 10 post(s)

Here's an article by Ellis that gives more detail on his preference for using beginning of year equity instead of end of year when calculating ROE:

AAII ROE Article

Doug

gerlach  
#9 Posted : Tuesday, November 11, 2008 5:24:55 AM(UTC)
gerlach

Rank: Administration

Posts: 437

Thanks: 7 times
Was thanked: 14 time(s) in 10 post(s)

Do you see the strange trend lines in any other companies, or in new companies that you download? Can you try downloading a new datafile for GILD and see if you still see the odd trend lines?

Doug

GeneBR  
#10 Posted : Tuesday, November 11, 2008 6:53:58 AM(UTC)
GeneBR

Rank: Advanced Member

Posts: 62

Was thanked: 1 time(s) in 1 post(s)






Doug, look at XTO.  Before you eliminate
outliers, trend lines looks about what you would expect.   Then knock
out 1998 thru 2001 to see a very skewed result.    Gene


heberling  
#11 Posted : Tuesday, November 11, 2008 7:18:26 AM(UTC)
heberling

Rank: Newbie

Posts: 3

Doug,

As you asked, I deleted my GILD ssg file and reloaded the data. The TK6 trend lines looked good at first, whether or not I deleted outliers, and the TK6 trends matched the TK5 trend lines. However when I saved the ssg file and reloaded it, the trend lines reverted to the old poor fits. They also reverted to the old poor fits if I clicked on refresh. Hope this gives some clues as to the problem.

I will study my other ssg files to see if the results are consistent with the GILD results.

Thanks for you patience on the Section 2B issue. I will study the material and the TK6 manual.

Paul Heberling
gerlach  
#12 Posted : Tuesday, November 11, 2008 7:38:57 AM(UTC)
gerlach

Rank: Administration

Posts: 437

Thanks: 7 times
Was thanked: 14 time(s) in 10 post(s)

OK, I'm now able to replicate the problem with the trend lines. The key was that it only affects saved companies, not new companies.

This issue appears to affect companies with negative earnings/pre-tax profit in the first year(s) of their 10-year histories and for which you eliminate years as outliers, BUT it only appears in companies that you've saved and then opened from your library. When you download these companies the first time, the trend lines are drawn as you expect.

I've verified this with our engineers and so they'll work on a fix for this for the next maintenance release.

Thanks for your help, everyone!

DG

Users browsing this topic
Guest (3)
Forum Jump  
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.

Powered by YAF.NET | YAF.NET © 2003-2024, Yet Another Forum.NET
This page was generated in 0.104 seconds.