Sensitivity Analysis To Drive Square Footage Adjustment

[Tom Woolford]

I'll use sensitivity to confirm other methods or tighten a range, but I find paired sales are best for Baths, garages, pools etc. About the only thing I use sensitivity for exclusively is porches etc.

 

[Mike Garrett RAA]

and what is left is an adjustment of $zzz

 

[Renee Healion]

Just wanted to say I always enjoy Valueseeker's thoughtful questions and the replies. The stats discussion with Leased Fee was really good to read too a while back. Came back to AF to find LIKE buttoning just a little weird so I wanted to say it outright.

 

[residentialguy]

I found that the better I got at appraising, the less I made quantitative adjustments. The less you make quantitative adjustments, the harder it is to find fault. My advice...don't make small adjustments that you can't fully and unequivocally support. There is no requirement to have a "tight range" using quantitative adjustments

 

[Terrel L. Shields]

the less I made quantitative adjustments

Math wise, making too many adjustments isn't going to be more accurate either. I rarely make more than two adjustments to a given improvement. Land is not an improvement and I do a dollar for dollar adjustment there as the very first step.

 

[Valueseeker]

Thanks Renee!

Ok well dont kill me on this (regretting this already). There are condition adjustments, above and location for busy roads as well as the ones shown.

 

[bnmappraisal]

Valueseeker, first off, Kudos for posting part of your grid; you've got more stones than me! (and probably the majority of others on here!)

The first thing that jumps out at me, though, is why no GLA adjustments on comps 2, 3, 4 & 5?
I typically don't make adjustments for less than ~100sf difference in GLA, but said comps are well beyond that and I would think generally would have an adjustment. There are exceptions, but even that being said, comps 2 & 5 are approximately 300 - 400 sf smaller; I would think the avg buyer would notice said difference (unless you're adjusting for BR's instead of GLA?)

 

[Amy Perkins]

That's interesting that you make the per sf adjustments last. I am more certain about those adjustments than the others; at least in my market. I found that external factors had to do with the market and would change. My reasoning is that in a seller's market, buyers are less picky. Conversely, in a buyer's market, buyers are more picky. This is why external obsolescence tends to be a riskier property.

 

[J Grant]

Thanks Renee!

Ok well dont kill me on this (regretting this already). There are condition adjustments, above and location for busy roads as well as the ones shown. View attachment 31282

Why are you using an excel spreadsheet giving you unreliable results? You posted in prior threads about statistics and regression re exploring it, but if you dont' understand how to use it or are using it with such weird results (of $2 per sf adjustment ), $2 per sf makes no sense in cost approach nor in price reaction seen for larger sf , so why use it? The sales most relevant to subject are there on your grid. Do a simple but effective sensitivity analysis on the grid with those sales.

Comp at 2060 sf is closest to subject 19060 sf. Look at the value difference of its adjusted price vs adjusted value of the comps.

You already adjusted for other factors of value and now are left with blank SF on the grid. I rough calculated by that grid contributory value per SF as anywhere from $20 to $25 per sf. Put in $20 per sf in software and see if it tightens up the value range, and then see if you want to refine it further for best results. Sensitivity analysis on the grid you adjust for other factors, the blank is then adjusted for amount that tightens the value range up ( and that makes sense )

 

[Gobears81]

In Central IL, for the price range that you are discussing, if you presented the choice of $4,000 or a two-car garage, the majority of prospective buyers would take the garage. Same with $5,000 and perhaps even a little higher. I understand the YMMV side, but couldn't resist mentioning that.

OP-I'm not sure what excel sheet is being circulated around here, but are you extracting the adjustments based on the sales used in the adjustment grid? If so, that would suggest inbreeding, which would result in all of your comps essentially adjusting to the same value. Have you tried using a separate set of comps for this type of analysis? I see you are using 5 comps in the report-are there additional sales that are comparable? Just remember that statistics is only one piece of the puzzle. It can (and will) drive erroneous conclusions in many cases.