Weighing Comp’s and “Kind-of-Comp’s”
The cornerstone of establishing market value for a given property is the CMA, or “Comparative Market Analysis.”
As prepared by both Realtors and Appraisers, it consists 
of selecting a peer group of three homes (“Comparable Sold Properties”), then going through a compare-and-contrast process sizing up each of the Comp’s relative to the “subject” home (the one you’re trying to price).
The pro’s call this “bracketing, post-adjustment.”
Integral to this process is the presumption that there are three good Comp’s: homes similar in style, size, condition, and location — that have sold within the last six months (preferably less).
Which Criteria to Relax
The art of valuing a given property comes in when the Comp “pickings,” so to speak, are slim.
Which criteria do you relax?
To take just one example, suppose you’re trying to price a rambler (Midwestern for “one-story,” called a “ranch” elsewhere).
Unfortunately, you have to go back 18 months to find a similar rambler that sold in the same neighborhood.
Some Realtors and Appraisers will widen their geographic scope, and look outside the neighborhood in adjacent ones to find a more recent sale.
Given how discrepant even adjacent neighborhoods can be, I 
find that less reliable than staying within the same neighborhood, and going back further in time for a similar sale.
Tweaking the Settings
If you go that route, though, Step #2 is figuring out what prices have done, overall, in that neighborhood, during that interval.
Fortunately, MLS statistics readily allow such statistical interpolation.
Alternatively, I will “jump” housing styles to find a good (or acceptable) Comp when necessary.
So, normally I’d prefer an “apples-to-apples,” rambler-to-rambler comparison.
However, if that’s not possible, I’ll use a 1 1/2 or two-story, then use “above ground finished square feet” to compare the “Kind-of-Comp” to the subject property.
