- Find a Real Estate Professional
- Realtors®
- Mortgage Lenders
- Home Improvement Pros
- Other Real Estate Services
- Review an Agent, Lender or Pro
- Marketing on Zillow
- Real Estate Agent Advertising
- Join the Professional Directory
- Popular
- Real Estate Market Reports
- More
Answers (21)
- Pasadenan
- Contributions:21466
"have what Opinion/View, they should go away..." -
No one on this thread said anyone should "go away"... the only thing I was trying to point out is that there are ways to make the available data and available estimates "useful" for a variety of purposes, but that it is extremely easy to misinterpret what is provided if one is not careful.
No one on this thread said anyone should "go away"... the only thing I was trying to point out is that there are ways to make the available data and available estimates "useful" for a variety of purposes, but that it is extremely easy to misinterpret what is provided if one is not careful.
- sunnyview
- Contributions:25139
Sorry for the typo in my earlier post. I meant to say to make sure your HOME (not "hoe") facts were correct. Mavis Beacon will probably send me a nasty email about that one. No offense intended just my bad typing at work :P
- Tug of War
- Contributions:1970
"I can't think of any reason why someone would care about misleading numbers or numbers that can't be directly applied to their intended usage."
"To me its a waste of time... I guess they feel less valuable as human beings if they are told their home is worth less than they thought..."
Not understanding or getting why someone is no the Criteria as far as I know for who can Post, have what Opinion/View, they should go away...
If it's used for that reason or as a justification of that type of behavior then those using it/expressing it are just Flat Wrong and far more of a Threat to what this Forum is suppose to be than a Bus load of "He who shall not be named" posters"
Thatsa my Opinion
"To me its a waste of time... I guess they feel less valuable as human beings if they are told their home is worth less than they thought..."
Not understanding or getting why someone is no the Criteria as far as I know for who can Post, have what Opinion/View, they should go away...
If it's used for that reason or as a justification of that type of behavior then those using it/expressing it are just Flat Wrong and far more of a Threat to what this Forum is suppose to be than a Bus load of "He who shall not be named" posters"
Thatsa my Opinion
- Pasadenan
- Contributions:21466
I can think of a lot of reasons a person may care about getting useful numbers or reasonable numbers. I can't think of any reason why someone would care about misleading numbers or numbers that can't be directly applied to their intended usage.
Tax liability for an eventual sale may be one reason. Estimating how much one can refinance for in order to do improvements, repairs or put a child through college may be another reason.
Rough calculation of net worth for discussions with a potential spouse or potential divorce may be another reason.
Estimates of inheritance tax based on potential government legislative decisions may be another reason.
Or perhaps advanced estate planning issues for how property can be divided between eventual heirs...
Or maybe just re-diversification of assets for portfolio balancing.
In any case, random fluctuations and wide tolerance margins are not helpful for those purposes, thus one must look for approaches to getting numbers that are more useful for the intended purpose. And obviously, no one wants to pay for a new appraisal each month to track market changes.
Tax liability for an eventual sale may be one reason. Estimating how much one can refinance for in order to do improvements, repairs or put a child through college may be another reason.
Rough calculation of net worth for discussions with a potential spouse or potential divorce may be another reason.
Estimates of inheritance tax based on potential government legislative decisions may be another reason.
Or perhaps advanced estate planning issues for how property can be divided between eventual heirs...
Or maybe just re-diversification of assets for portfolio balancing.
In any case, random fluctuations and wide tolerance margins are not helpful for those purposes, thus one must look for approaches to getting numbers that are more useful for the intended purpose. And obviously, no one wants to pay for a new appraisal each month to track market changes.
- Tug of War
- Contributions:1970
Complain about the Hate and then show them the Love..lol
- Vince Curtis, "SoCal Appraiser"
- Contributions:4699
Pasadenan - Sorry, and you are right, I just dont get why people give a shXX about an opinion from an algorithim that changes daily....To me its a waste of time... I guess they feel less valuable as human beings if they are told their home is worth less than they thought...
- Pasadenan
- Contributions:21466
I disagree -- the question was asked in 2007 and never answered (except for REA Ted Mackel's response that his perspective is that Zillow is right less than 5% of the time) until I answered it on this thread.
The person got more frustrated with having a second property that seemed to have the same issue, but still no answer.
I believe I answered the question thoroughly, but I'm sure that if I didn't the poster should be back with more clarifications on the details of the question.
The person got more frustrated with having a second property that seemed to have the same issue, but still no answer.
I believe I answered the question thoroughly, but I'm sure that if I didn't the poster should be back with more clarifications on the details of the question.
- Vince Curtis, "SoCal Appraiser"
- Contributions:4699
seems to be lots of posters on here trying to show how clever they are - shame no one answered deaniewinters question
There is no question on this post, its a statement of Zillow-hate...
There is no question on this post, its a statement of Zillow-hate...
- Su2z
- Contributions:1
Here in CT, especially on the Shoreline, Zillow's Zestimates are so on target it would be creepy, if it wasn't so helpful. I'm grateful anyway. Just felt like sharing....
- hpvanc
- Contributions:2579
Deaniewinter,
Can you comfortable share any more specifics on the issue? I'm always happy to help flag what is clearly erroneous data so that sites like these can accomplish the good.
Can you comfortable share any more specifics on the issue? I'm always happy to help flag what is clearly erroneous data so that sites like these can accomplish the good.
- sunnyview
- Contributions:25139
Zillow has more on Zesitmates here and tools for owners here. Making sure you hoe facts are correct is one of the best ways to make sure that your Zestimate is as accurate as possible within the obvious constraints of an AVM like Zillow.
I would also check back in April when the owners estimate feature is supposed to come back to finish adding any remodels/additions that Zillow is not valuing.
I would also check back in April when the owners estimate feature is supposed to come back to finish adding any remodels/additions that Zillow is not valuing.
- Vince Curtis, "SoCal Appraiser"
- Contributions:4699
The site has a problem here and has had from the very beginning. I am sorry if someone doesn't want to hear about it.
I agree ! Ban the machines ! Appraisers only - we need the work !
I agree ! Ban the machines ! Appraisers only - we need the work !
- hpvanc
- Contributions:2579
Pasa,
How about backing off a little and let Deaniewinter have a chance to explain what the error is. You don't know that it is the property they have a Make Me Move price on that they are referencing, or if it is the property in question what might be in error about it.
If they just don't like the indicator that values have fallen, then you may be justified in what you have written.
How about backing off a little and let Deaniewinter have a chance to explain what the error is. You don't know that it is the property they have a Make Me Move price on that they are referencing, or if it is the property in question what might be in error about it.
If they just don't like the indicator that values have fallen, then you may be justified in what you have written.
- Pasadenan
- Contributions:21466
"Fix the PROBLEM Please" -
Obviously, I can't fix the problem as I don't work for Zillow; I'm just a data freak that found out about a good source for data at no cost to me. Of course one needs to be aware of potential data errors when using the data.
And I do know that the 5 statisticians are hard at work to try to address all known problems. I have had Email conversations with a few of them on various subjects of interest to me.
"I am sorry if someone doesn't want to hear about it." -
The statisticians rarely read the forum as they are too busy trying to get the work done. The customer service reps do regularly read the board, but most of them really don't have much connection to the math underlying the estimates, and they can only give the generic response that have been provided for them. But they are really good about changing an incorrect sales price really quickly if it is brought to their attention. Customer service is a different job then data analysis, which is also a different job than programming. (Only 5 programmers for the entire site presently, which are not the same as the 5 statisticians that do the data analysis, estimates, market indexes, and mortgage trends).
By the way, I'm also sorry if people really don't want to hear about the limitations of present technology. There are hundreds of AVM programs and sites available out there, all competing for limited market share, but the present state of the art just is not capable of doing what many keep thinking is promised. The competition is good as it inspires innovation to get better numbers faster and closer to market value than their competitors. And they are always open to new innovative ideas, but the concept of "just fix it" is not very helpful.
Obviously, I can't fix the problem as I don't work for Zillow; I'm just a data freak that found out about a good source for data at no cost to me. Of course one needs to be aware of potential data errors when using the data.
And I do know that the 5 statisticians are hard at work to try to address all known problems. I have had Email conversations with a few of them on various subjects of interest to me.
"I am sorry if someone doesn't want to hear about it." -
The statisticians rarely read the forum as they are too busy trying to get the work done. The customer service reps do regularly read the board, but most of them really don't have much connection to the math underlying the estimates, and they can only give the generic response that have been provided for them. But they are really good about changing an incorrect sales price really quickly if it is brought to their attention. Customer service is a different job then data analysis, which is also a different job than programming. (Only 5 programmers for the entire site presently, which are not the same as the 5 statisticians that do the data analysis, estimates, market indexes, and mortgage trends).
By the way, I'm also sorry if people really don't want to hear about the limitations of present technology. There are hundreds of AVM programs and sites available out there, all competing for limited market share, but the present state of the art just is not capable of doing what many keep thinking is promised. The competition is good as it inspires innovation to get better numbers faster and closer to market value than their competitors. And they are always open to new innovative ideas, but the concept of "just fix it" is not very helpful.
- Pasadenan
- Contributions:21466
To get the Zestimate monthly data in an Excel file format from January 2002 forward:
Use the URL:
http://www.zillow.com/ajax/homedetail/HomeValueChartData.htm?mt=1&zpid=16440799&format=json
(Change the ZPID number (Zillow Property Identification Number) as appropriate for the specific property of interest, as indicated on the URL for the home details page, as well as on the home details page.)
You will need to save the file, and then change the },{ to "enter"s. You can do it with some word processors. Notepad will let you do it if you do the search for },{ and press "escape", and then "enter" F3, repeating "enter" F3 until all are done.
You can then input as coma delimited text into excel.
The date is in modified UNIX milisecond time. Thus, divide by 1000, divide by 3600, divide by 24, then add 12/31/1969 5:00 pm. (different hour for desired time zone).
This is not something Zillow has sanctioned for public use; just some underlying data I happened to find, that I think most people should have access to for their further analysis. I have provided this information in several other locations on this forum so far, and so far, it has not been deleted nor have I been told to stop telling people how to do it.
This is the data used to generate the charts. They did change the format recently to improve the loading time for the charts, so at least one of my posts on this subject have an outdated method of processing the data.
(I found the URL for the data using Firefox's adblock plus add-on).
Use the URL:
http://www.zillow.com/ajax/homedetail/HomeValueChartData.htm?mt=1&zpid=16440799&format=json
(Change the ZPID number (Zillow Property Identification Number) as appropriate for the specific property of interest, as indicated on the URL for the home details page, as well as on the home details page.)
You will need to save the file, and then change the },{ to "enter"s. You can do it with some word processors. Notepad will let you do it if you do the search for },{ and press "escape", and then "enter" F3, repeating "enter" F3 until all are done.
You can then input as coma delimited text into excel.
The date is in modified UNIX milisecond time. Thus, divide by 1000, divide by 3600, divide by 24, then add 12/31/1969 5:00 pm. (different hour for desired time zone).
This is not something Zillow has sanctioned for public use; just some underlying data I happened to find, that I think most people should have access to for their further analysis. I have provided this information in several other locations on this forum so far, and so far, it has not been deleted nor have I been told to stop telling people how to do it.
This is the data used to generate the charts. They did change the format recently to improve the loading time for the charts, so at least one of my posts on this subject have an outdated method of processing the data.
(I found the URL for the data using Firefox's adblock plus add-on).
- Pasadenan
- Contributions:21466
Semi Valley property purchased 2/1/08, $815k:
Presently "make me move" at $950k;
present estimate method when purchased: $901k
Bubble peaked: $1.1 Million, December 2006.
All values from January 2006 forward were recalculated in June, 2011, due to a change in estimate modeling method that improved the tolerance range of the estimates over-all (not individually) by about 50%. All values will be recalculated from January 1996 forward in the near future when some of the known issues have been worked out, and when the present learning models have learned some of the market response behavior for the over 1 million different computer models.
It is true that that "great buy" compared to previous Feb 2008 Z-estimate is not as good a buy when compared to the present calculation method. But Zillow is not liable for people misusing their estimates, and their are plenty of disclaimers, and the tolerance range has been published from the beginning.
If you were reading the forum from 2007 through 2008, you would know full well that the market was still bubbled, and still in decline, and would have extrapolated from the 1997 value instead.
Median value for Semi Valley homes is now $378k (August). This means the home in question is a "high end" home, and thus shouldn't be compared to the median trend, but the upper tier trend. One should normalize the upper tier trend for January 2006 forward by taking the average of the estimates and upper tier trend, taking the ratio, then multiplying. This will smooth out the fluctuations known to occur in individual estimates due to small sample size of what sells is a short time period in a small region.
Here is the link to the upper tier trend:
Simi Valley 10 yr upper tier trend
There is a button on that page for downloading the data in Excel format.
Here is the upper tier trend chart:
Presently "make me move" at $950k;
present estimate method when purchased: $901k
Bubble peaked: $1.1 Million, December 2006.
All values from January 2006 forward were recalculated in June, 2011, due to a change in estimate modeling method that improved the tolerance range of the estimates over-all (not individually) by about 50%. All values will be recalculated from January 1996 forward in the near future when some of the known issues have been worked out, and when the present learning models have learned some of the market response behavior for the over 1 million different computer models.
It is true that that "great buy" compared to previous Feb 2008 Z-estimate is not as good a buy when compared to the present calculation method. But Zillow is not liable for people misusing their estimates, and their are plenty of disclaimers, and the tolerance range has been published from the beginning.
If you were reading the forum from 2007 through 2008, you would know full well that the market was still bubbled, and still in decline, and would have extrapolated from the 1997 value instead.
Median value for Semi Valley homes is now $378k (August). This means the home in question is a "high end" home, and thus shouldn't be compared to the median trend, but the upper tier trend. One should normalize the upper tier trend for January 2006 forward by taking the average of the estimates and upper tier trend, taking the ratio, then multiplying. This will smooth out the fluctuations known to occur in individual estimates due to small sample size of what sells is a short time period in a small region.
Here is the link to the upper tier trend:
Simi Valley 10 yr upper tier trend
There is a button on that page for downloading the data in Excel format.
Here is the upper tier trend chart:
- Pasadenan
- Contributions:21466
From 2007:
"Something is incorrect with your estimator tool. When you have a house estimate for over $633,000 and another house 3 houses down which is an exact same model house and in better condition estimate for 541,000 it doesn't compute. It is a 92,000 difference. How does that happen?" -
It happens because people don't pay the same amount, nor buy on the same date. The machine method assumes if someone pays more for the "same thing", it must not be the same thing, it must have better views, or better finishes, or better amenities, or higher ceilings, or more insulation..., and if the market was increasing and someone paid more later, it might not be sufficiently more to account for the market changes. The learning modeling compensates for the market changes using market index factors similar to what is indicated on the local info pages (tab above).
Also, the machines have absolutely no method of seeing nor adjusting for deterioration of the property since last purchased, or improvements to the property since last purchased.
"Something is incorrect with your estimator tool. When you have a house estimate for over $633,000 and another house 3 houses down which is an exact same model house and in better condition estimate for 541,000 it doesn't compute. It is a 92,000 difference. How does that happen?" -
It happens because people don't pay the same amount, nor buy on the same date. The machine method assumes if someone pays more for the "same thing", it must not be the same thing, it must have better views, or better finishes, or better amenities, or higher ceilings, or more insulation..., and if the market was increasing and someone paid more later, it might not be sufficiently more to account for the market changes. The learning modeling compensates for the market changes using market index factors similar to what is indicated on the local info pages (tab above).
Also, the machines have absolutely no method of seeing nor adjusting for deterioration of the property since last purchased, or improvements to the property since last purchased.
- Pasadenan
- Contributions:21466
"...is to question the human being who is responsible for entering incorrect information" -
I don't question anyone who inputs wrong information; we all do that...
Just go to the home details page(s) where the wrong information is, and use the "report a problem" button under the edit pull down tab under the main photo and be specific about the data that needs correction. An "estimate" is not "data", it is a machine opinion.
And if there was a huge spike or huge drop in the estimate, just note that so that someone else can look into the recent sales in the area, unless of course you can identify the recently sold in the area that has either a wrong sold price or wrong sqft, or wrong lot size....
The estimates are highly dependent on the last sold price adjusted for last sold date, and the method used is known to have wide fluctuations about the nominal value... that is a "feature" to respond rapidly to market changes and to converge quickly on the nominal value. Most people don't understand this as they expect values to go up slightly with the consumer price index each year. But that is not really how most markets behave, especially when the market is correcting for bubble speculation, and compounded with high unemployment.
I don't question anyone who inputs wrong information; we all do that...
Just go to the home details page(s) where the wrong information is, and use the "report a problem" button under the edit pull down tab under the main photo and be specific about the data that needs correction. An "estimate" is not "data", it is a machine opinion.
And if there was a huge spike or huge drop in the estimate, just note that so that someone else can look into the recent sales in the area, unless of course you can identify the recently sold in the area that has either a wrong sold price or wrong sqft, or wrong lot size....
The estimates are highly dependent on the last sold price adjusted for last sold date, and the method used is known to have wide fluctuations about the nominal value... that is a "feature" to respond rapidly to market changes and to converge quickly on the nominal value. Most people don't understand this as they expect values to go up slightly with the consumer price index each year. But that is not really how most markets behave, especially when the market is correcting for bubble speculation, and compounded with high unemployment.
- hpvanc
- Contributions:2579
Can you comfortable share any more specifics on the issue? I'm always happy to help flag what is clearly erroneous data so that sites like these can accomplish the good.
Hopefully some of the other regular posters feel the same way instead of just jumping on the 'Zestimates are great your just trying to manipulate them' or "Zestimates need to be removed immediately,' fight that seems to be going on in the forums right now.
Hopefully some of the other regular posters feel the same way instead of just jumping on the 'Zestimates are great your just trying to manipulate them' or "Zestimates need to be removed immediately,' fight that seems to be going on in the forums right now.
- deaniewinter
- Contributions:3
I guess someone doesn 't like to hear the truth about a real problem. It's always easier to look the other way than it is to find a solution.
The answer to a machine generated problem is to question the human being who is responsbile for enterting incorrect information. The site has a problem here and has had from the very beginning. I am sorry if someone doesn't want to hear about it.
I believe this site has the potential to cause great harm as well as great good. Someone needs to be a good watch dog! Fix the PROBLEM Please.
- Pasadenan
- Contributions:21466
There is nothing "wrong" with the Zillow estimator...
it says right here:
Zestimate Tolerance Range
That 15% of the estimates will be more than 20% different than present market value in Ventura County, and there is absolutely nothing anyone can do about it.
Pretty soon, no one will believe anything people that complain about machine generated statistics post.
it says right here:
Zestimate Tolerance Range
That 15% of the estimates will be more than 20% different than present market value in Ventura County, and there is absolutely nothing anyone can do about it.
Pretty soon, no one will believe anything people that complain about machine generated statistics post.
Why is your estimator so screwed up, pretty soon no one will believe anything this site has to say!!
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.