Okay, now we're getting somewhere.... this property is an "over Improvement" for the development not functional obsolescence. I would certainly do an analysis of the entire development (time does not matter... research all the properties three to five years back if you have to) to make sure that your subject property is the white elephant of the development.
I disagree. You need to go to different, alternative, competitive areas to make sure that a 1700 sq ft dwelling is an over Improvement.... if you find one, grid it and do a location adjustment. Otherwise, you won't be bracketed.
Also, from the alternative, competitive neighborhoods with 1700 square foot properties, you'll be able to derive a percentage difference between 1,000 and 1700 square foot sales. With this percentage difference, you won't need to use linear regression.