Satistics linear regression concept question?

From the choices below, the best reason for not making estimates of the response for values of the predictor variable that are outside of the valid prediction range is:

a)The predictor variable is undefined

b)The characteristics of the model may change

c)There is no rationale, estimates can be made anywhere the predictor variable is defined

d)Points outside the VPR are outliers

e)Points outside the VPR are too far away from (x,y)

f)The response variable is undefined

Thanks for the help!!!