Solved

Excel 2003 to Excel 2010 LINEST formula

Posted on 2013-11-21
8
1,164 Views
Last Modified: 2013-12-04
My formula =INDEX(LINEST(M$21:M$23,$L$21:$L$23^{1,1},,TRUE),1,3)

Works great in 2003 but when I run the sheet in 2010 the value's do not populate correctly. Was anything change from 2003 to 2010 in this formula?
0
Comment
Question by:Keef4000
  • 3
  • 3
  • 2
8 Comments
 
LVL 50

Expert Comment

by:barry houdini
ID: 39667576
Seems to work OK for me - I don't think anything changed in Excel 2010 that would affect LINEST - what do you mean "the values do not populate correctly" - do you get errors or the wrong results?

regards, barry
0
 

Author Comment

by:Keef4000
ID: 39667608
I figured something out that is very interesting. So the formula works like you said. However, I have 3 formula's like the one above.

Excel 2003 (This is what its suppose to return)


=INDEX(LINEST(M$21:M$23,$L$21:$L$23^{1,1},,TRUE),1,1)
            Returns: 0
=INDEX(LINEST(M$21:M$23,$L$21:$L$23^{1,1},,TRUE),1,2)
            Returns: -0.000917151
=INDEX(LINEST(M$21:M$23,$L$21:$L$23^{1,1},,TRUE),1,3)
            Returns: 0.084003538


However, In Excel 2010

=INDEX(LINEST(M$21:M$23,$L$21:$L$23^{1,1},,TRUE),1,1)
            Returns:  -0.000917151
=INDEX(LINEST(M$21:M$23,$L$21:$L$23^{1,1},,TRUE),1,2)
            Returns: 0
=INDEX(LINEST(M$21:M$23,$L$21:$L$23^{1,1},,TRUE),1,3)
            Returns: 0.084003538


As you can tell the 1 formula and 2 formula are flipped in 2010. For kicks and giggles I reversed formula 1 and 2 on 2010 and My results are correct. I don't understand how that is possible. Example of my flip are below:

=INDEX(LINEST(M$21:M$23,$L$21:$L$23^{1,1},,TRUE),1,2)
            Returns: 0
=INDEX(LINEST(M$21:M$23,$L$21:$L$23^{1,1},,TRUE),1,1)
            Returns: -0.000917151
=INDEX(LINEST(M$21:M$23,$L$21:$L$23^{1,1},,TRUE),1,3)
            Returns: 0.084003538
0
 
LVL 50

Expert Comment

by:barry houdini
ID: 39667730
OK, I see what you mean but I have no idea why these are different - I don't have Excel 2003 but it seems that there is a difference even between Excel 2007 and Excel 2010 (and the issue is with LINEST not INDEX because I took out the latter and there are differences between what LINEST returns in both versions).

I'll see if anyone else can help out........

regards, barry
0
Live: Real-Time Solutions, Start Here

Receive instant 1:1 support from technology experts, using our real-time conversation and whiteboard interface. Your first 5 minutes are always free.

 
LVL 81

Accepted Solution

by:
byundt earned 500 total points
ID: 39668144
Microsoft was chastised by the academic community for errors in their statistical functions, including LINEST and LOGEST. As a result, they have revisited the code for those functions on several occasions, with the result being that you get different answers depending on which version you are using. In most cases in my experience, the differences are minor.

The most extensive rewrite occurred in Excel 2013, where a third-party firm with strong ties to academia compared the results of the revised algorithms to widely accepted statistical software packages, and found excellent agreement. In some cases, the new algorithm was noticeably slower than the original one--so the old version was retained and the new version got a different name.

Bottom line: this is a known problem. Description of the LINEST function in Excel 2003 and in later versions of Excel
0
 
LVL 81

Assisted Solution

by:byundt
byundt earned 500 total points
ID: 39669762
In your desired formula, what is the purpose for listing a duplicate set of independent variables?
=INDEX(LINEST(M$21:M$23,$L$21:$L$23^{1,1},,TRUE),1,3)
In your formula, the ^{1,1} produces two sets of L21:L23 as values of "x" variables. But they aren't independent, so one of the two sets should be thrown out of a proper correlation. That's what Excel 2013 does--throw out the first set of L21:L23 by returning 0 for the coefficient.

To get the constant term (y-intercept), I might have used:
=INDEX(LINEST(M$21:M$23,$L$21:$L$23,,TRUE),1,2)
0
 

Author Closing Comment

by:Keef4000
ID: 39677874
Thanks for the information. I think we will have to stay with 2003 then.
0
 
LVL 81

Expert Comment

by:byundt
ID: 39677908
I think we will have to stay with 2003 then
Ooh! That's the wrong takeaway from Microsoft's KnowledgeBase article. They have finally fixed the problems with LINEST and other statistical functions in Excel 2013. Clinging to Excel 2003 will only perpetuate the possibility of error.
0
 

Author Comment

by:Keef4000
ID: 39696519
Thank you for the recommendation I will revisit 2013 then. We have highly statistical data that has to return consistent variable throughout many different condition so the numbers have to be correct or we cannot ship parts out the door. We have considered writing our own formula as well which might be the direction to go.
0

Featured Post

Courses: Start Training Online With Pros, Today

Brush up on the basics or master the advanced techniques required to earn essential industry certifications, with Courses. Enroll in a course and start learning today. Training topics range from Android App Dev to the Xen Virtualization Platform.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

This article will show you how to use shortcut menus in the Access run-time environment.
Do you use a spreadsheet like Microsoft's Excel?  Have you ever wanted to link out to a non excel file on your computer or network drive?  This is the way I found to do it!
This Micro Tutorial demonstrate the bugs in Microsoft Excel for Mac with Pivot Charts.
Polish reports in Access so they look terrific. Take yourself to another level. Equations, Back Color, Alternate Back Color. Write easy VBA Code. Tighten space to use less pages. Launch report from a menu, considering criteria only when it is filled…

785 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question