Accuracy of Quadratic Regression

[bernhardSA]

Dear Thread Responders

I appreciate the quick and insightful answers. This has been very useful. The first set of "less accurate" results were via the Statistics2Var app. I have been able to get the same high accuracy results with CAS command line programming, and the modified x-vector in the statistics app as suggested.

It is a little surprising that HP chose to segregate the lower precision for HOME and higher precision for CAS. Why not simply do all the math at highest precision and display in most accurate floating point format or super snazzy symbolic/fraction/exact format as required by user. A high school student may not care (or appreciate HOME vs CAS subtlety) - they just need to get the same answer as the teacher's solution key, and using the statistics app may seem faster more attractive than explicitly typing in matrices and commands ... it stands to reason that they might expect the same answer regardless of the path.

As a longstanding old school HP user and fan, and one who was trying to convert his high school freshman child to HP and RPN, this has been an unexpected speedbump.

Again thx for the quick help (I will keep the HP and my high schooler will now get a TI)