Voice of the Appraiser In-Depth Report

[George Hatch]

There is no such thing as an assumption-free appraisal. And although USPAP doesn't explicitly address the criteria for the use of standard assumptions there are minimum criteria which apply to the use of EAs, which TBH are but a subset of all assumptions that appraisers use. "Because I say so" isn't enough and to be sure, that isn't an ultimatum that appraisers are being faced with.

Here's what's required for the use of an EA (which is different in concept to a normal assumption only because it's specific to a particular assignment).


So that will raise the question "what comprises a reasonable basis to assume that info is sufficient accurate for use in an appraisal? " Another way to put that question is "what reason do I have to distrust the accuracy of this info?" Sadie just mentioned reviews where the reviewer can see why they might have reason to distrust the accuracy of a specific individual. It seems to me that the same level or more scrutiny and consideration would be applicable to the review and analysis of the info in a PDR.

People obviously disagree as to how much to trust the PDRs. But I think it would be unreasonable to assume that the GSEs themselves have not been making the comparisons between the performance of the PDRs vs appraisers. In any case they are apparently demonstrating their conclusions by their actions.

I think most appraisers would be loathe to consider the possibility that the combination of these PDRs + the AI interpretation software could possibly equal or conceivably even exceed the performance of the appraisers. If that's not already the case then it could possibly occur at some point in the near or more distant future.

No matter what, using a single AI app to identify specific makes/models of appliances and fixtures and floor coverings and ceiling heights and overall condition and such has one inherent advantage over the combined performance of appraisers: As a singular app it would at least be internally consistent even if the accuracy is suboptimal. Maybe not already at that point, but sooner or later that evolution must logically be considered inevitable.

 

[TerryRohrer]

I don't disagree with the above, but must ask what "accuracy" means. In the video of Lyle Radke the other day, he mentioned "accuracy" then quickly qualified that with "that is hard to measure." We see a lot of statements regarding the accuracy of this and the accuracy of that and particularly the accuracy of condition and quality ratings and the accuracy of appraisals. Yet, no one has defined that term in those contexts nor, more importantly, how it is being measured in order to draw the stated conclusion. If accuracy is assumed to have the usual meaning of being "correct or precise" then how is everyone measuring it? I expect from the GSE position, accuracy is judged by any difference between an appraiser and the GSE's position on anything (which we know is pliable).

 

[George Hatch]

I never really thought about it before, but maybe appraisers might have use for running an AI app on their end to add a layer of consistency to their quality/condition ratings. At least then they could point to how they came to their ratings.

Perhaps not even necessary to take that step every time. Just often enough to provide feedback to the appraiser for how they're doing compared to the app. A means to calibrate the appraiser's judgement and performance.

Alternately, an appraiser doing these desktops already has access to the pics in the PDR along with any ratings in the PDR. They can agree or disagree as necessary but sooner or later they're going to see enough examples of how the PDRs are rating to calibrate their own judgement and to apply those to what they're seeing in their own inspections and among the MLS listings they're analyzing in their SC analysis.

 

[Dublin ohio]

And to starve you out.

One could only hope

 

[TerryRohrer]

The problem for appraisers, I think, is that the GSEs are demanding consistency between appraisers regarding quality and condition ratings in their misguided/mistaken/ignorant effort to model all things all the time everywhere with the press of a button. It is not about an appraiser's valuation. It is about data collection, or in the case of these two variables, data creation.

 

[djd09]

cant be much lower then appraisers rooting for other appraisers to starve out...but it is the same propaganda they been spewing for years

 

[George Hatch]

cant be much lower then appraisers rooting for other appraisers to starve out...but it is the same propaganda they been spewing for years

You're the one who thinks they are out to get you. I was just feeding your troll. For some reason you seem to object, which is weird because my feeding your boundless paranoia should be making you happier.

 

[djd09]

cant be much lower then appraisers rooting for other appraisers to starve out...

 

[djd09]

i mean really...it is not me calling other appraisers racist like TAF and the unethical stakeholders...that aint paranoia

 

[djd09]

now go hit the report button...you big baby