Run a similar analysis and identify whether the FP added $1000 or $2500 in contributory value in each individual sale? Garages $3000, $5000 or $7500 per car bay? What about GLA, how much variation from each sale to the next? Be exact, as that is what you seem to allude is required for any adjustment.
OK, lets do as you ask using the same data set.
Instead of the figures I presented being a SP:LP percentage lets say they represent the extracted surplus GLA adjustment $/sf of the 84 sales. If you will recall, the data set had a range of 92 to 121 with a median of 103 so these would be $/sf in this exercise. In my market this would be fairly typical $/sf range for surplus GLA of a 2,000 sq ft home worth around $700,000.
Lets say the subject is 2,000 sq ft and one of the comps is 2,200 sq ft. If I use the median adjustment, $103/sf in this example, my adjustment is $20,600 but even at the most historical extreme of $121/sf the adjustment is only $24,200 or a maximum error in the adjusted price of $3,800. No error is desirable but a $3,800 error to one comparable on a $700,000 property probably does not have a substantial impact on the credibility of the appraisal report for most intended uses.
On the contrary, an 18% error made when applying a SP;LP in this same example of a $700,000 home is an 18% error in the adjusted price, or $126,000. There are not many instances that I am aware of where a $126,000 error in the adjusted price of a comparable would not have an effect on the credibility of the report of a $700,000 property.
Same data and same potential error rate but the application of one has a dramatically different effect on the adjusted price than the other. This is only one reason I dislike SP:LP adjustments.