The minus value at the end of the equation is perhaps not a coefficient but the intercept of the equation after the application of independent variable coefficients. This can very well be negative as the coefficients alone result in an inflated sales price that must be adjusted by this amount for the correct value when using the equation for prediction. If this is from the independent analysis for the stories variable then it relates to only that variable when compared to the sales and is not applicable to the overall multiple regression equation. Lets say stories is 2 and the sales price is 105,000 then a coefficient of 65,000 would make 130,000 and need to deduct 25,000 to get the correct sales price.
