Getting coefficients of a quadratic equation based on three points

David Aranovsky wrote: Thanks Stuart, that looks like it worked.

Good. I checked it in Prime 1 after I posted and it worked as well.

But then the question follows: why does MathCad expect you to guess values for variables that we trying to solve for? Is that just for the sake of declaration? If that is the case, then it seems awkward, especially, since these values are arbitrary.

As Fred says, the numerical solvers generally require a starting point (guess values) - the solvers are fairly general and a number of problems have solutions that are dependent on the initial values.

Local minima and maxima are a good example - starting on side of the 'hill' or the other may take you to completely different minima. I posted a worksheet on the Hooke-Jeeves search algorithm here: http://communities.ptc.com/message/185455#185455 . It contains an example near the bottom of what happens when you specify different initial values.

Stuart

This is actually a set of linear equations in the unknowns, so here are a couple of ways of solving them without the need for initial guesses:

Alan