How many comps do you use?

[Zoe]

Bracketing all these items should not enter one’s mind when selecting comparables for a property.

I get why it’s done, the worst thing you can have is a revision request, they’re time-consuming, completely kill production and you’re not paid for it, so sometimes it’s easier just to throw in a bull**** sale that isn’t really a comparable. But at the same time, you should never put a sale in the grid that you can’t defend as being truly comparable. And because it brackets the garage count doesn’t mean it’s comparable

Yes, your and my focus should always be on the subject and definition of value and property rights being appraised. You don't need comps that don't compete with the subject. It is fine to expand the data for market trends and comparison purposes but you want the best comps and most similar. You want the ones that are most competitive with the subject.

In a homogeneous market, no big deal. In a heterogeneous like I work most of the time, it is a big deal. If an appraiser don't know the market they are working, they can get in trouble.

I don't see how appraisers work in markets they don't know. Like in different States and such or even in same State.

I can see it if subject is located in a very homogeneous subdivision, but all subjects are not that way.

 

[Surf Cat]

These are the old school ways though. Now, it's more about data collection, AI, and what a bunch of data being crunched says..... the independent art of it is going away.

In many cases, using grouped data or some other type of analysis is far better than just using one sale that happens to "bracket" a feature.

See?

 

[DWiley]

See?

It is not "old school" to use unsupported adjustments. Its just wrong.

 

[Zoe]

Tennessee overall is very heterogeneous market. I'll break it down. Rural is very heterogeneous. Knoxville may be most homogeneous market over all as a major city. Nashville is probably next but I have worked Nashville on multifamily. Nashville is not as heterogeneous as Memphis but more homogeneous than Knoxville. Memphis urban and suburban is very heterogenous. If you don't know what you are doing, you could get in trouble as an appraiser.

Memphis is very segmented market for the subject in many many cases.

some might say complex vs non-complex.

 

[Amy Perkins]

I’ve seen this question get posed some times over the years, but was wondering how many comps you residential appraises are typically using in your market, and what market that is. I live in a unique area where I have urban, suburban, and rural coverage areas. When I started some years ago my mentor was notorious for using tons of comps, like 6-7 sales then 2-3 actives/pendings. When I joined my current firm they encouraged me to cut back on using unnecessary comps when they don’t strengthen the credibility and aren’t an assignment condition (e.g 2 actives). Over the years I’ve widdled down my comps and have found myself often using 3-4 sales only on conventional, particularly in easier markets or rural markets with few comps where I can only find about three good ones. I often feel guilty when I send out a 3 comp report despite it being completely credible and bracketing features etc. there’s still some that end up with perhaps 5-6 sales 1-3 listings, typically the more unique ones, but overall find myself using only 3-4 for conventional 95% of the time. I’d like to hear everyone else’s thoughts and experiences. It’s certainly saved me a lot of time and have had 0 pushback, and always focus on keeping it credible regardless

I start with 10-12 and then I eliminate the ones I dont like.

 

[Amy Perkins]

I agree with this approach. Over the years I have seen a lot of reports where adding more comps made the report worse, not better.

Think about it this scenario - you have 3 or 4 comps that support a value between $450,000 and $500,000, but then you add 2 or 3 other sales and the adjusted range widens to $400,000 to $600,000. Is that a "better" report? With that scenario the reviewer will often start to wonder adjustment(s) (or adjustment rate) you missed that resulted in such a wide range.

Lol, I typically never use 3-4 unless there really some reason why like no sales. My comps are typically eliminated when it suggests a wider range. This is just how I determine value, adding more sales just to add more sales it not my aim. Instead I am trying to narrow down value.

 

[Amy Perkins]

Bracketing comes into play sometimes. Sometimes you can't bracket something without using something that don't even compete with the subject. Usually an underwriter or reviewer will realize that. But some don't really understand appraisal.

I try to bracket or provide a matched pair.

 

[TerryRohrer]

See?

So, your main point is that you don't have any idea what "most probable" means? Did your mentor ever tell you that one sale doesn't make a market? If one sale doesn't make a market, how could one sale possibly reflect "the" likely impact of a single feature on prices? Those appraisers who steadfastly refuse to learn, or even just understand, "anything" about data analysis or statistics, even though myriad sources outside of appraisal have been developing the field for years, are why requests for these data are being developed. As a whole, appraisers learn what questions will be asked when their reports are delivered. Then, they deliver reports with the sole intent of insuring the fewest number of those questions will be asked by dumping garbage in their reports. Take any issue, and it happens. And dumping one sale in a report and saying, "see, covered it" is just another way of dumping the garbage. Actual analysis, on the other hand, can actually support the conclusion reached by the appraiser.

The following is an excerpt from the last residential report I delivered. Every aspect of the subject was "bracketed" by the 5 sales discussed. A focus on "bracketing" removes focus from "analyzing."

"the prices of these sales all fall within a narrow range (the 5 most heavily weighted had a coefficient of variation (COV) of about 5%), but their property characteristics varied widely. The following are characteristics commonly found to influence sale prices of residential properties, with the COV of that characteristic (in parentheses) for these 5 sales: Age (44%); GLA (30%); Finished basement (75%); Unfinished basement (32%); Baths (24%); Garage stalls (87%)."

 

[Zoe]

Some might say cookie cutter vs non-cookie cutter.

 

[Amy Perkins]

Some might say cookie cutter vs non-cookie cutter.

I never get cookies anymore