How many comps do you use?

[Surf Cat]

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

I see what you did there DW.... you got all excited thinking you were a field appraiser again. Then you came to your Freddie Mac senses and had to subliminally interject the non-subject inspecting, non- delineating the neighborhood, non-curbside comparable viewing, chained to the desk, data crunching appraiser. Because that's where all this is going.

Don't worry, I believe most here would do the same thing in your shoes.

 

[DWiley]

I see what you did there DW.... you got all excited thinking you were a field appraiser again. Then you came to your Freddie Mac senses and had to subliminally interject the non-subject inspecting, non- delineating the neighborhood, non-curbside comparable viewing, chained to the desk, data crunching appraiser. Because that's where all this is going.

When the appraiser applies adjustments for a feature, the support required for that adjustment is not dependent on the type of appraisal (desktop, hybrid, traditional). And, as previously noted, the requirement to actually have that support is not a new thing. That is not "where this is all going." It has what has always been required.

 

[alebrewer]

I typically stick to only one comp in an SCA. Easier to estimate adjustments.

 

[Surf Cat]

When the appraiser applies adjustments for a feature, the support required for that adjustment is not dependent on the type of appraisal (desktop, hybrid, traditional). And, as previously noted, the requirement to actually have that support is not a new thing. That is not "where this is all going." It has what has always been required.

Well, first off, the subject matter of how many comps do you use, segwayed into bracketing. You're now talking about adjustments and support for those adjustments no matter which type of assignment. Of course support for adjustments is not a new thing. That said, GSE lenders have been accepting reports with non-supported adjustments for decades..... everyone knows this.

In regards to the subject matter, I laid out everything I do in post #3 of this thread. That methodology has served me well. Does it work on every assignment? Of course not, real estate is imperfect. There's always anomalies and Oddball assignments.

The independent fee appraiser going out into the field, scouring the MLS, reading each MLS listing, communicating with agents to see how it matches up with the subject property......is going away.

As you say; "using grouped data or some other type of analysis is far better than just using one sale that happens to "bracket" a feature" is the new fashion trend.

Whether I agree with that or not, that's where this is going....Grouped data, AI picking comps, no pesky commentary.
The opinion of value just has to be "good enough"....or hit the purchase price to remain on the AMC roster.

The lenders need to make loans to make money. They don't need the old school, speed bump appraiser in the way. Crunch the numbers, turn the report, done deal..... without even going and starting the car.

 

[J Grant]

As you say; "using grouped data or some other type of analysis is far better than just using one sale that happens to "bracket" a feature".

Is using grouped data or some other type of analysis far better than an appraiser who has spent hours vetting sales, to find 2, or 3, or 4 good comps that bracket a feature and displays the extraction and adjustment right there on the grid seen by a reader, inlcuding sucinct narrative that makes sense to the reader?

Why are appraisers who are crappy at bracketing , which is a fundamental skill, hired? If that same clueless appraiser uses grouped data or some other type of analysis, they will still do a crappy job, but it will be far more difficult to see it.

The advantage of traditional appraisal techniques is that they mirror the behaviour of actual buyers, filtered through the appraisal development. Actual buyers speak with their wallets WRT what they will pay for positive features or the discount they expect for adverse defects. It is not rocket science. What takes time is searching for the right comps to show it.

 

[Surf Cat]

Is using grouped data or some other type of analysis far better than an appraiser who has spent hours vetting sales, to find 2, or 3, or 4 good comps that bracket a feature and displays the extraction and adjustment right there on the grid

It is now....

 

[DWiley]

Is using grouped data or some other type of analysis far better than an appraiser who has spent hours vetting sales, to find 2, or 3, or 4 good comps that bracket a feature and displays the extraction and adjustment right there on the grid seen by a reader, inlcuding sucinct narrative that makes sense to the reader?

The answer to that question is objectively "yes." If one uses 4 comps, and only 1 or 2 or those comps has the feature, then there are only 1 or 2 data points supporting the adjustment. Supporting an adjustment with more data is always better. That is not a new thing either. And, by the way, group data is a "traditional appraisal technique." I was using that method 30 years ago.

 

[ZZGAMAZZ]

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

YEA but you just described both sides of the double edged sword....

 

[DWiley]

Is using grouped data or some other type of analysis far better than an appraiser who has spent hours vetting sales, to find 2, or 3, or 4 good comps that bracket a feature and displays the extraction and adjustment right there on the grid seen by a reader, inlcuding sucinct narrative that makes sense to the reader?

The answer to that question is objectively "yes." If one uses 4 comps, and only 1 or 2 or those comps has the feature, then there are only 1 or 2 data points supporting the adjustment. Supporting an adjustment with more data is always better. That is not a new thing either. And, by the way, group data is a "traditional appraisal technique." I was using that method 30 years ago.

 

[J Grant]

The answer to that question is objectively "yes." If one uses 4 comps, and only 1 or 2 or those comps has the feature, then there are only 1 or 2 data points supporting the adjustment. Supporting an adjustment with more data is always better. That is not a new thing either. And, by the way, group data is a "traditional appraisal technique." I was using that method 30 years ago.

Grouped data can add additional support. Why would it be a first, or only method? Skipping bracketing with a good comp , (Usually several comps ) is suspect unless there are none available.
One or two or (or more) good comps can take hours to find after eliminating 10 other sales,. These comps are not "data points. " They are carefully vetted indicators of value.

Buyers do not spend 600k to buy a "data point." A comp presents a house that is an alternative choice to a buyer.

The star of the URAR show is the sales comparison approach, not grouped data.