RCA
Elite Member
- Joined
- Jun 27, 2017
- Professional Status
- Certified General Appraiser
- State
- California
From a very objective point of view, and having talked to a few people high up, there is a view that the AVMs are going to be improved through various methods.
I assume a method similar to RCA will be used to generate a good model for properties in different market areas, residuals calculated, comparables ranked by residual, getting an objective ranking of properties by subjective factors such as quality, condition, and appeal, and then scoring the properties by their ranking. Then, the MLS photos of the comparables will be given to AI to train it to score unsold properties based on their photos. We might expect some standardization of property photos along the way. So, in this way, the AVM can objectively score a subject property based on its photos in comparison to those of the comparables. No appraiser is needed in most cases, - so it is hoped.
Non-parametric statistics and data mining will be used to provide fairly accurate appraisals in, it is expected, something like 90% of appraisals. Exceptionally superior or inferior properties, or properties in rural areas with few comps, will require an appraiser on an as-needed basis. To fully implement this will take 5-10 or so years.
Then there will be a competitive fight for market share between AMCs. Many will disappear. -- But so will many appraisers.
I assume a method similar to RCA will be used to generate a good model for properties in different market areas, residuals calculated, comparables ranked by residual, getting an objective ranking of properties by subjective factors such as quality, condition, and appeal, and then scoring the properties by their ranking. Then, the MLS photos of the comparables will be given to AI to train it to score unsold properties based on their photos. We might expect some standardization of property photos along the way. So, in this way, the AVM can objectively score a subject property based on its photos in comparison to those of the comparables. No appraiser is needed in most cases, - so it is hoped.
Non-parametric statistics and data mining will be used to provide fairly accurate appraisals in, it is expected, something like 90% of appraisals. Exceptionally superior or inferior properties, or properties in rural areas with few comps, will require an appraiser on an as-needed basis. To fully implement this will take 5-10 or so years.
Then there will be a competitive fight for market share between AMCs. Many will disappear. -- But so will many appraisers.