What if you test all coefficients in your model and determine that none of the independent variables are statisically significant (and the F-test varifies this), does this simply mean you should scrap the model and try new explanatory variables?
What’s the next route?