- Find a Real Estate Professional
- Alabama Realtors®
- Alabama Mortgage Lenders
- Alabama Home Improvement Pros
- Alabama Other Real Estate Services
- Review an Agent, Lender or Pro
- Marketing on Zillow
- Real Estate Agent Advertising
- Join the Professional Directory
- More
Answers (7)
- hpvanc
- Contributions:2579
http://www.zillow.com/wikipages/What-is-a-Zestimate/
Subitron,
To add a hyperlink, highlight text already in your post and then click the hyperlink icon. It does work, it is just quirky.
Subitron,
To add a hyperlink, highlight text already in your post and then click the hyperlink icon. It does work, it is just quirky.
- Pasadenan
- Contributions:21467
By the way, I don't believe I was condescending of anyone, nor claiming any unusual skills; only making the observation that the proprietary method used is not easily understood by most people even if "trade secrets" are disclosed; and Zillow has no intention of disclosing their trade secrets nor proprietary method.
But if one really wished to pursue it more, I would suggest reading U.S Patent #5361201
owned by First American Core Logic,
that expires next year and becomes "public domain" so that anyone can use it without royalties.
Since First American Core Logic sued 12 AVM companies for alleged patent infringement of that patent including Zillow to try to discover the methods of their competitors by using the court "disclosure" process, it is pretty clear that Core Logic believes their competitors are doing something right and may have better estimates in some cases.
And clearly, Zillow's method has some similarities to the referenced patent since the Zillow Statisticians have made it clear they do not use linear regression for their learning modeling method.
So, if one can understand what some refer to as "fuzzy math"; by all means, read up on the patent and do some of your own tests to verify for yourself what happens when one uses these kinds of approaches to track market changes.
But if one really wished to pursue it more, I would suggest reading U.S Patent #5361201
owned by First American Core Logic,
that expires next year and becomes "public domain" so that anyone can use it without royalties.
Since First American Core Logic sued 12 AVM companies for alleged patent infringement of that patent including Zillow to try to discover the methods of their competitors by using the court "disclosure" process, it is pretty clear that Core Logic believes their competitors are doing something right and may have better estimates in some cases.
And clearly, Zillow's method has some similarities to the referenced patent since the Zillow Statisticians have made it clear they do not use linear regression for their learning modeling method.
So, if one can understand what some refer to as "fuzzy math"; by all means, read up on the patent and do some of your own tests to verify for yourself what happens when one uses these kinds of approaches to track market changes.
- Pasadenan
- Contributions:21467
"AND INCORRECT aerial views of their respective properties" -
Aerial views are not used at all in the estimating method, but only based on an approximate Geocode extrapolated from the U.S. Postal Service 9 digit Zip code geo code database, and referenced to images by Pictometry International Corporation via MicroSoft Bing's website.
There is presently no exact method to obtain a geocode. The U.S. Census Bureau stated they would be geocoding all residential units in the U.S. for the 2010 Census. I have no clue if they did this or not, but if they did, it will eventually improve the geocoding for sites like this as public dollars spent on such projects makes it public data.
If one is not happy with the "fuzzy math" used on this website; there are plenty of other websites that use different "fuzzy math", such as:
Homesand.net
Eppraisal
realquest.com express
cyberhomes.com
trulia
chase
homegain
homes.com
Aerial views are not used at all in the estimating method, but only based on an approximate Geocode extrapolated from the U.S. Postal Service 9 digit Zip code geo code database, and referenced to images by Pictometry International Corporation via MicroSoft Bing's website.
There is presently no exact method to obtain a geocode. The U.S. Census Bureau stated they would be geocoding all residential units in the U.S. for the 2010 Census. I have no clue if they did this or not, but if they did, it will eventually improve the geocoding for sites like this as public dollars spent on such projects makes it public data.
If one is not happy with the "fuzzy math" used on this website; there are plenty of other websites that use different "fuzzy math", such as:
Homesand.net
Eppraisal
realquest.com express
cyberhomes.com
trulia
chase
homegain
homes.com
- Hamp Yonce, "Zilluminati"
- Contributions:3463
Dr Spock was very busy when he was forced to correct the peon's illogical beliefs regarding illogical-ness. :-)
- Tug of War
- Contributions:1984
"You wouldn't understand the explanation regardless. No one else that asks that question seems to understand the explanation."
Just a bit Condescending
A lot of "You wouldn't understand" people out there" guess they ain't as mentally advanced as you either
The Fuzzy Math of Home Values
I'm sure the Poster is sorry they annoyed you
Just a bit Condescending
A lot of "You wouldn't understand" people out there" guess they ain't as mentally advanced as you either
The Fuzzy Math of Home Values
I'm sure the Poster is sorry they annoyed you
- wetdawgs
- Contributions:26854
These are the data I see for the three properties you mention. Are there data errors in the public records? You win for Zestimate $/sq ft by a huge margin. (I do not represent Zillow)
- Pasadenan
- Contributions:21467
You wouldn't understand the explanation regardless. No one else that asks that question seems to understand the explanation.
The method uses modeling of recently sold non-distressed ownership housing units in the area to 9 county record variables:
1) house sqft
2) lot sqft
3) # bedrooms
4) # bathrooms
5) # stories
6) year built
7) tax assessed value (adjusted for local rules)
8) last sold price extrapolated from sold date
9) covered parking (if in the county records).
Then those recently sold are checked against the model, and if way off, they are removed from the model and the model recalculated.
The model is then used for extrapolating machine opinions of approximate value. Each opinion is calculated entirely independently with absolutely no regard for the neighboring properties. There is a wide tolerance margin for each estimate due to small numbers that sell in a limited time span in a limited area. But the smaller areas and smaller time spans are still chosen to track the market better, even though it does mean more random fluctuations and wider tolerance margins.
Since the fluctuations are entirely independent and within the tolerance margins, only if you look at a long period of time can you see that the general trends of the property and neighboring properties are similar.
Since it is all automated; no one pays any attention to the details on why one goes up and another goes down for any given estimate cycle.
Sure, it is relatively easy to plot the estimate trends of the properties in question on the same chart; but since I had a hard drive die, I'm not going to do that at this time. But you could do it yourself. The URL to get the data is:
http://www.zillow.com/ajax/homedetail/HomeValueChartData.htm?mt=1&zpid=#&format=json
where the # is replaced with the Zillow property identification number as indicated on the home details page and also used in the web address of the home details page.
If you do that, the date is in milliseconds from December 31, 1969. (Modified Unix time)
The method uses modeling of recently sold non-distressed ownership housing units in the area to 9 county record variables:
1) house sqft
2) lot sqft
3) # bedrooms
4) # bathrooms
5) # stories
6) year built
7) tax assessed value (adjusted for local rules)
8) last sold price extrapolated from sold date
9) covered parking (if in the county records).
Then those recently sold are checked against the model, and if way off, they are removed from the model and the model recalculated.
The model is then used for extrapolating machine opinions of approximate value. Each opinion is calculated entirely independently with absolutely no regard for the neighboring properties. There is a wide tolerance margin for each estimate due to small numbers that sell in a limited time span in a limited area. But the smaller areas and smaller time spans are still chosen to track the market better, even though it does mean more random fluctuations and wider tolerance margins.
Since the fluctuations are entirely independent and within the tolerance margins, only if you look at a long period of time can you see that the general trends of the property and neighboring properties are similar.
Since it is all automated; no one pays any attention to the details on why one goes up and another goes down for any given estimate cycle.
Sure, it is relatively easy to plot the estimate trends of the properties in question on the same chart; but since I had a hard drive die, I'm not going to do that at this time. But you could do it yourself. The URL to get the data is:
http://www.zillow.com/ajax/homedetail/HomeValueChartData.htm?mt=1&zpid=#&format=json
where the # is replaced with the Zillow property identification number as indicated on the home details page and also used in the web address of the home details page.
If you do that, the date is in milliseconds from December 31, 1969. (Modified Unix time)
your values are totally illogical! - Please explain!
Multiple dwellings in my complex which were all built at the same time (like 37 and 39 Canterbury Way) are priced $15K+ more than my residence at 1 Canterbury Way and have LESS bedrooms, ZERO upgrades, less square footage/acreage AND INCORRECT aerial views of their respective properties !!
PLEASE explain this to me and why home value keeps getting downgraded and not at the same pace as the 2 aforementioned properties????
Stating a discriminatory preference in an advertisement for housing is illegal. If you think this content is discriminatory or otherwise inappropriate and feel it should be removed from Zillow, please let us know by completing the information above.
We will review this content. Thanks for helping make the site more useful to everyone. To learn more, read Zillow's Good Neighbor Policy.