View previous topic :: View next topic |
Author |
Message |
mslan
Joined: 23 Apr 2007 Posts: 2
|
Posted: Thu Apr 26, 2007 8:51 am Post subject: Basketball and MiniTab |
|
|
Hey,
Whats going on everyone?
I was just wondering if anyone was familiar with a stats program called MiniTab and if anyone has any ideas if this can be helpful to me in building mutliple regressions to attempt to create my own predictions on basketball stats?
If anyone here has done anything relating to basketball within this program could you please lend a helping hand?
Thanks a lot. |
|
Back to top |
|
|
Ryan J. Parker
Joined: 23 Mar 2007 Posts: 711 Location: Raleigh, NC
|
Posted: Thu Apr 26, 2007 9:53 am Post subject: |
|
|
I've never used it, so I can only suggest R (it's free). |
|
Back to top |
|
|
tpryan
Joined: 11 Feb 2005 Posts: 100
|
Posted: Sat Apr 28, 2007 1:30 am Post subject: |
|
|
First of all, it is "Minitab". I have used it in teaching for almost 30 years and am a member of the Author Assistance Program of Minitab, Inc. since I have written statistics books that they list as Minitab Companion Texts.
It will do just about anything in multiple regression that you would want to do.
R, which was suggested, is best suited for specialized methods not available in the best known and most popular statistical software.
For routine work in multiple regression, Minitab is far superior. |
|
Back to top |
|
|
Ryan J. Parker
Joined: 23 Mar 2007 Posts: 711 Location: Raleigh, NC
|
Posted: Sat Apr 28, 2007 6:53 am Post subject: |
|
|
Well I'd hope so... at $1,200 per single user license. |
|
Back to top |
|
|
ChrisH
Joined: 02 Sep 2006 Posts: 4 Location: Chicago
|
Posted: Mon Apr 30, 2007 9:02 am Post subject: |
|
|
you can get a student copy for a lot less. or at least you could a couple years ago. |
|
Back to top |
|
|
AJax
Joined: 22 Jan 2006 Posts: 49 Location: The Prairies of the Great Middle West
|
Posted: Mon Apr 30, 2007 11:46 am Post subject: |
|
|
Minitab is $122 for students at my university. Stata is comparably priced, and better in my opinion. It's also cross-platform, and I think Minitab only works on windows. R is cool and open-source but it's more for the specialist.
A lot of things presented on this forum like figuring correlations can be done with spreadsheets. For the next bump up in complexity - least squares and logistic regression, model comparison - something like Minitab would be great; SPSS, Stata, and SAS would work just as well too. |
|
Back to top |
|
|
tpryan
Joined: 11 Feb 2005 Posts: 100
|
Posted: Mon Apr 30, 2007 12:07 pm Post subject: |
|
|
Minitab has a 30-day free trial. Many of my Internet short course students have utilized the free trial as it fits the time frame for the course. |
|
Back to top |
|
|
jkubatko
Joined: 05 Jan 2005 Posts: 702 Location: Columbus, OH
|
Posted: Mon Apr 30, 2007 1:27 pm Post subject: |
|
|
tpryan wrote: | For routine work in multiple regression, Minitab is far superior. |
Far superior? Or far easier? Big difference. _________________ Regards,
Justin Kubatko
Basketball-Reference.com |
|
Back to top |
|
|
tpryan
Joined: 11 Feb 2005 Posts: 100
|
Posted: Mon Apr 30, 2007 9:07 pm Post subject: |
|
|
jkubatko wrote: | tpryan wrote: | For routine work in multiple regression, Minitab is far superior. |
Far superior? Or far easier? Big difference. |
In my opinion, the primary drawback to R (and S-Plus) for the casual user is that it is necessary to write code to do simple things. As far as I'm concerned, that shouldn't be necessary.
R is inadequate for the casual user; Sandy Weisberg indicates problems with R for such users at http://www.stat.umn.edu/~sandy/UseR06.pdf.
I will undoubtedly use R eventually in working on the 2nd edition of my regression book, but only for the specialized stuff. (I have used S-Plus in the past for regression work.)
Minitab is not great for "modern" regression, but most users do not even use regression software capability that existed 20 years ago!
It is easy to write Minitab macros (I have written a few dozen) and the library of Minitab macros at Minitab, Inc. can be downloaded from there. Macros don't help on computer-intensive stuff, however, because the code in macros must be interpreted rather than compiled. So there is no hope of doing something like LTS (least trimmed sum of squares) efficiently in Minitab unless it is provided in a future release and Minitab, Inc. has not moved into areas such as robust regression.
So R and S-Plus easily outclass Minitab in certain areas of regression, but Minitab has good regression capabilities for the vast majority of users and uses. |
|
Back to top |
|
|
jkubatko
Joined: 05 Jan 2005 Posts: 702 Location: Columbus, OH
|
Posted: Mon Apr 30, 2007 10:00 pm Post subject: |
|
|
Thanks for the clarification, Tom. I think you have to be careful about using words like "far superior." Many people would mistake that as a statement about the accuracy of Minitab as compared to R. _________________ Regards,
Justin Kubatko
Basketball-Reference.com |
|
Back to top |
|
|
tpryan
Joined: 11 Feb 2005 Posts: 100
|
Posted: Tue May 01, 2007 1:41 pm Post subject: |
|
|
jkubatko wrote: | Thanks for the clarification, Tom. I think you have to be careful about using words like "far superior." Many people would mistake that as a statement about the accuracy of Minitab as compared to R. |
Accuracy generally isn't a problem for most statistical procedures with today's algorithms for virtually any statistical software product. Accuracy was a problem decades ago, however, and datasets such as the Longley (1967) dataset were used to test the accuracy of regression output. |
|
Back to top |
|
|
jkubatko
Joined: 05 Jan 2005 Posts: 702 Location: Columbus, OH
|
Posted: Tue May 01, 2007 3:03 pm Post subject: |
|
|
tpryan wrote: | Accuracy generally isn't a problem for most statistical procedures with today's algorithms for virtually any statistical software product. Accuracy was a problem decades ago, however, and datasets such as the Longley (1967) dataset were used to test the accuracy of regression output. |
I know that and of course you know that, but I doubt the general public does. _________________ Regards,
Justin Kubatko
Basketball-Reference.com |
|
Back to top |
|
|
mslan
Joined: 23 Apr 2007 Posts: 2
|
Posted: Wed May 02, 2007 6:34 pm Post subject: |
|
|
thanks a lot for all your advice gentlemen. |
|
Back to top |
|
|
|