Unless these regressions are nonlinear in the unknown parameters, then SOLVER
is a very inefficient and potentially inaccurate way to do this.

For simple linear regressions, consider the worksheet functions SLOPE,
INTERCEPT, RSQ, and STEYX

For multiple regression, consider the worksheet functions LINEST or LOGEST.

Jerry

"Jasper" wrote:

Hey there everybody,

while designing a statistical procedure, I was tempted to use Excel 97, for
convinience of extreme oversight. However, I need to solve 16,000
regressions. Not terribelly hard stuff, though the VBA code I designed does
not return any value, or keep the value. This is the code I use, can someone
see where things go wrong? Thanks in advance!

Sub test()
SolverReset
SolverOk SetCell:="$DV$4, MaxMinVal:=2, ValueOf:=1", ByChange:="$DV$4"
SolverSolve UserFinish:=False
'Changing userfinish from false to true does not yield any difference
SolverFinish KeepFinal:=1
'Changing this line into SolverFinish KeepFinal:=1, ReportArray:=1, will
only yield an error on the lack of availability of memory
End Sub