Appraisal Statistics: Regression - Do Your Really Know How Much Work It Is?

[RCA]

"Understanding this is not really exact..." sounds a lot like the appraisal process but I appreciate the way you try to simplify the process. The software naturally has come a long way since I studied this years ago - I feel like I'm trying to understand a modern programming language from a punch-card perspective!

I briefly read the heteroskedasticity section and felt myself being pulled again by the dark side and, after all these years blissfully honing my human intelligence at the expense of keeping up with artificial intelligence, I'm going to have to walk away from this discussion to make some more money. There's a drug to this topic, and as a wonk myself I can feel the serotonin brain-bath dulling my perspective. "How'd your day go, hon?" Well, I didn't make any money but I really enjoyed getting a glympse of MLS data in four dimensions again..."

Good luck, Bert! I'm more interested in the policies and safeguards of AI in real estate valuation than the application these days. I'll leave you with the thought that just because we "can" doesn't mean that we "should" from an economical perspective. I hope in your continued R&D you do not stop arguing for Mr. Boots, the appraiser-on-the-ground to be part of the process. It would be a shame to "prune" the human intelligence and find that it really was the best way after all.

MARS sometimes starts off arbitrarily splitting data sets, because initially it can't see any useful relationships to make that first split. So it in fact just splits it somewhere and then starts go forward improving things. In the end it winds up with a split at the root level that make sense - just like I said in terms of how people think. That is the human factor -we can look at something complex and somehow make sense of it. But for the computer - it has to do a lot of its own footwork to get to that point.

Yea, we all have to make money. ...

 

[KYLECODY]

I had a realtor call me one time on a flip I was selling. He was trying to justify his low asking price based on his "regression analysis". He had taken a class and had all these interesting things/theories he was basing his ask price on. It sounded like Latin to me at the time and lacked common sense versus comparing that listing to closed, similar sales in that subdivision. It reminded me of some of the stuff in this thread. Its an intellectual conversation but may not pass the real world logic test.

 

[Carnivore]

12345 - not useful content

 

[Michael P Jacobs MAI]

I'm of the opinion residential appraisers should avoid trying to mimick avms altogether and be better at the basics.

I agree, Joe, but AVMs are here to stay and where is the voice that can counter Big Data/Artificial Intelligence ("AI") promise of faster-better-cheaper? No, appraisers are not "making buggywhips" as some of them suggest, but most of us - including me - are only able to "speak" buggywhip. But these guys are in love with the process - it's a religion after you've been on the treadmill a while - and so they don't want to hear about the advantages of boots-on-the-ground because that's just negative talk. OUR profession is just a reminder that that they haven't yet been "worthy" of the grace necessary to crack the code. We need the counter-spell for this magic; its necessary to convey our message in their language.

Why? Because the people who "sell" this stuff to Wall Street are very good at promising the world and obfuscating the discrepancies ("I was there"). The upper-management decision-maker that is so removed from the street due diligence process "will be sold" the simple-solution every time because there's no alternative voice and they don't have the background to question the math. Everyone they are surrounded by falls in love with the wonkish promises and lines up to buy-in because it all seems so tidy - the bean-counters love the cost savings, Quality Assurance loves that "it's not their fault anymore," and the commissioned (or bonused) rain-makers LOVE that there's so little friction shoveling money out of the vault.

And then Mr. Boots-on-the-Ground appraiser occasionally reads a story about all this happening to the profession, raises his or her head from the forms and computer screen, and sputters, "...but, but, but a machine can never replace a human..." They respond, "I like you very much but if that's all you've got we'll have to try the Big Data/AI solution."

I really don't know what the solution is, but the absence of any experienced AND educated appraiser's voice is troubling while the Big Data/AI marketing department continues to encroach on the (MY) profession.

 

[Joe Flacco]

I agree, Joe, but AVMs are here to stay and where is the voice that can counter Big Data/Artificial Intelligence ("AI") promise of faster-better-cheaper? No, appraisers are not "making buggywhips" as some of them suggest, but most of us - including me - are only able to "speak" buggywhip. But these guys are in love with the process - it's a religion after you've been on the treadmill a while - and so they don't want to hear about the advantages of boots-on-the-ground because that's just negative talk. OUR profession is just a reminder that that they haven't yet been "worthy" of the grace necessary to crack the code. We need the counter-spell for this magic; its necessary to convey our message in their language.

Why? Because the people who "sell" this stuff to Wall Street are very good at promising the world and obfuscating the discrepancies ("I was there"). The upper-management decision-maker that is so removed from the street due diligence process "will be sold" the simple-solution every time because there's no alternative voice and they don't have the background to question the math. Everyone they are surrounded by falls in love with the wonkish promises and lines up to buy-in because it all seems so tidy - the bean-counters love the cost savings, Quality Assurance loves that "it's not their fault anymore," and the commissioned (or bonused) rain-makers LOVE that there's so little friction shoveling money out of the vault.

And then Mr. Boots-on-the-Ground appraiser occasionally reads a story about all this happening to the profession, raises his or her head from the forms and computer screen, and sputters, "...but, but, but a machine can never replace a human..." They respond, "I like you very much but if that's all you've got we'll have to try the Big Data/AI solution."

I really don't know what the solution is, but the absence of any experienced AND educated appraiser's voice is troubling while the Big Data/AI marketing department continues to encroach on the (MY) profession.

It is a complicated issue. I think in residential appraisal many are using fannie mae guidelines as a how-to-appraise manual. For example draw a one mile radius and then match up GLA and bedroom, bathroom counts to select comps. This type of process is easily replicated by a machine which I think ends up making AVM's seem more reliable than they are compared to a properly developed appraisal.

Besides that, we can never be faster or cheaper than AVMs so there can no question that the appraisal is better. The main problem with residential appraisals is too much generic boilerplate and not enough info specific to the assignment. Both appraisal and AVM gives you a value but only appraisal can explain exactly why the opinion of value is what it is. Anybody should be able to pick up the report and understand why the value is what it is and be convinced. This is the one advantage we have so I think that's where we have to improve.

I appreciate that Bert is trying to make a case for Mr. Boots. I didn't know that was what he was trying to do. We can show them the math but in the end I think the appraisal report needs to speak for itself.

 

[RCA]

It is a complicated issue. I think in residential appraisal many are using fannie mae guidelines as a how-to-appraise manual. For example draw a one mile radius and then match up GLA and bedroom, bathroom counts to select comps. This type of process is easily replicated by a machine which I think ends up making AVM's seem more reliable than they are compared to a properly developed appraisal.

Besides that, we can never be faster or cheaper than AVMs so there can no question that the appraisal is better. The main problem with residential appraisals is too much generic boilerplate and not enough info specific to the assignment. Both appraisal and AVM gives you a value but only appraisal can explain exactly why the opinion of value is what it is. Anybody should be able to pick up the report and understand why the value is what it is and be convinced. This is the one advantage we have so I think that's where we have to improve.

I appreciate that Bert is trying to make a case for Mr. Boots. I didn't know that was what he was trying to do. We can show them the math but in the end I think the appraisal report needs to speak for itself.

Hybrid appraisals are based on an inspection done by a separate property inspector, whom the appraiser is expected to trust. IF, the AVM companies like HouseCanary ever sufficiently train such property inspectors to do accurate measurements of property per ANSI/BOM standards, to correctly classify property condition, quality, view, properly catch and note defects, draw floorplans, and stand behind the results, then most of that work will disappear. What is left is filling in the cracks, investigative work and tying it all together. I don't know why current home inspectors couldn't take on the task of also being property inspectors. I think the two have to be merged into one profession, - especially if property inspectors don't have to be concerned with valuation. It seems quite doable.

Then, if and when that happens, appraisers won't be "boots on the ground" any longer. But that is all a long ways off. If you want a better future, don't want to deal with statistics, pick up training as a home inspector, study construction, go that route. At least you will still have the fun of driving around to see the world, rather than be stuck all day at some desk. People who want stay in valuation should extend in the direction of learning to value more kinds of property types - such as equipment appraisal. That is to say the inspection part of appraisal is probably going to merge with home inspection, given time. Valuation will require more general expertise that goes way beyond just appraising a single family residence. In fact, I think equipment appraisal is probably going to grow rapidly - with rapidly changing technology. Robots will compound the rate of change.

I have to wonder whether companies like House Canary think they can eventually contract property inspection to property inspectors with E&O, pass the results on to appraisers with their own E&O who pay a hefty fee to use their online tools, then by controlling the money stream for the whole process take their "fair" share and not have to worry about liability. That might well be their long term plan. ... Well life goes on.

 

[Joe Flacco]

Hybrid appraisals are based on an inspection done by a separate property inspector, whom the appraiser is expected to trust. IF, the AVM companies like HouseCanary ever sufficiently train such property inspectors to do accurate measurements of property per ANSI/BOM standards, to correctly classify property condition, quality, view, properly catch and note defects, draw floorplans, and stand behind the results, then most of that work will disappear. What is left is filling in the cracks, investigative work and tying it all together. I don't know why current home inspectors couldn't take on the task of also being property inspectors. I think the two have to be merged into one profession, - especially if property inspectors don't have to be concerned with valuation. It seems quite doable.

Then, if and when that happens, appraisers won't be "boots on the ground" any longer. But that is all a long ways off. If you want a better future, don't want to deal with statistics, pick up training as a home inspector, study construction, go that route. At least you will still have the fun of driving around to see the world, rather than be stuck all day at some desk. People who want stay in valuation should extend in the direction of learning to value more kinds of property types - such as equipment appraisal. That is to say the inspection part of appraisal is probably going to merge with home inspection, given time. Valuation will require more general expertise that goes way beyond just appraising a single family residence. In fact, I think equipment appraisal is probably going to grow rapidly - with rapidly changing technology. Robots will compound the rate of change.

Thanks for the advice?

If you have not been reading my comments, I have been saying statistics is less reliable than old school methods done properly. Let me clarify any confusion. I'm saying F statistics. Also, they will never get home inspectors on board because they are trying to pay pennies.

 

[RCA]

Thanks for the advice?

If you have not been reading my comments, I have been saying statistics is less reliable than old school methods done properly. Let me clarify any confusion. I'm saying F statistics. Also, they will never get home inspectors on board because they are trying to pay pennies.

Well I could give you a dataset like the classic Boston Housing Dataset and see where you get. I don't think you would take the challenge. No on does. No one can compete with MARS for extracting adjustments on a dataset like this. But you are free to try. BTW "medv" is the median price of each census tract (like a sales price). Give me a predicted price (either medv or cmedv) on all of the census tracts using your method. With MARS I can do that, with an R2 of over 90%. BTW, this same dataset is frequently used in statistics competitions.

[Likewise, you can run a MARS regression going back many years on all the sales transactions in a neighborhood to pick up nice adjustments on home features that just won't be possible if you focus on the last 12 months.]

Anyway, the original data are 506 observations on 14 variables, medv being the target variable:

crim: per capita crime rate by town
zn: proportion of residential land zoned for lots over 25,000 sq.ft
indus: proportion of non-retail business acres per town
chas: Charles River dummy variable (= 1 if tract bounds river; 0 otherwise)
nox: nitric oxides concentration (parts per 10 million)
rm: average number of rooms per dwelling
age: proportion of owner-occupied units built prior to 1940
dis: weighted distances to five Boston employment centres
rad: index of accessibility to radial highways
tax: full-value property-tax rate per USD 10,000
ptratio: pupil-teacher ratio by town
b: 1000(B - 0.63)^2 where B is the proportion of blacks by town
lstat: percentage of lower status of the population
medv: median value of owner-occupied homes in USD 1000's

The corrected data set has the following additional columns:

cmedv: corrected median value of owner-occupied homes in USD 1000's
town: name of town
tract: census tract
lon: longitude of census tract
lat: latitude of census tract

http://lib.stat.cmu.edu/datasets/boston_corrected.txt

 

[Joe Flacco]

Data for 80% of factors and differences for residential real estate is not even collected anywhere. What and how differences impact a certain property depends on the property. I p iss on your MARS. It sounds to me like you do not understand residential real estate.

 

[residentialguy]

I p iss on your MARS.

Oh boy....wait till Joan hears about this!