[olecapt]

Originally Posted by Terri B

Well, I got luck, the ones I did find were very similar to mine. Have done tons of research at the assessor's site. Really tough finding anything selling in the month of 1007 and 0907. I am just a homeowner trying to make my way through all of this.

Drop me a DM with the address and I will run you up a quick market analysis. The way the market is RE Agent have lots of time for pro bomo comps.

I am sure Eric would do the same if you prefer.

[Eric Young]

Originally Posted by olecapt

Drop me a DM with the address and I will run you up a quick market analysis. The way the market is RE Agent have lots of time for pro bomo comps.

I am sure Eric would do the same if you prefer.

I was thinking the same thing, OleCapt. It's 5:30 and I'm already done for the day.

[ParkTwain]

Originally Posted by Eric Young

In the case of the property that I pulled from Cyberhomes, it failed to list as a comp a very similar home within two hundred yards that closed in September. My guess is that the algorithm ignored it because it was 200 SF larger than the target property. However, it was especially relevant because it was sold for a substantial lower price than an identical match that sold several months earlier.

I don't hear you talking about true outliers. They do happen. One happened in my neighborhood last year, something like a 50K to 75K spike, or maybe it was more. I just remember that it got my attention as being outside the previously established selling range for that size home in my neighborhood. People make some kind of special deal. The owner dies, and it's picked up by distant relatives, etc, etc.

What do you think the software should do about them? Recognize that the sell price is some number of standard deviations away from a N-month moving average for that property size, therefore ignore it?

There are lots of variables in homes to consider. There are relatively few direct comps for any home on the market. I'm not interested in a RE agent pushing me toward a particular price because of outliers. The trend downward (in the current environment) must already be established distinctly from any outliers. One or two sales don't make a trend.

[Eric Young]

Originally Posted by ParkTwain

I don't hear you talking about true outliers. They do happen. One happened in my neighborhood last year, something like a 50K to 75K spike, or maybe it was more. I just remember that it got my attention as being outside the previously established selling range for that size home in my neighborhood. People make some kind of special deal. The owner dies, and it's picked up by distant relatives, etc, etc.

What do you think the software should do about them? Recognize that the sell price is some number of standard deviations away from a N-month moving average for that property size, therefore ignore it?

There are lots of variables in homes to consider. There are relatively few direct comps for any home on the market. I'm not interested in a RE agent pushing me toward a particular price because of outliers. The trend downward (in the current environment) must already be established distinctly from any outliers. One or two sales don't make a trend.

PT -

You're right - I'm not talking about true outliers at all. I would, however, suggest that the software may not be capabable of picking up a non-arms length transaction.

When I see a sale price that is substantially different than the expected norm, I'll look to see if the buyer and seller share a last name or address, or if the transaction was probated, etc. While these factors are available in the public record, I would be surprised if the software designers are incorporating them much less be able to recognize that a sale by John and Mary Doe to the JMDoe, LLC is a non-arm's length transaction.

Further, listings usually have agent remarks as well as multiple photos that my human eyes can include in a final determination of likely selling price (which I do on outliers as well as those within the expected norm). Also available on the MLS and of particular importance is pricing history which shows a property's original listing price and each change in the listing price as well as the date. Studying price histories provides an excellent view of the strength of the market in a particular neighborhood and also allows me to easily filter out the fraudulent sales (original price is X - no sale for four months - price raised to X + $50,000 - sale in one day at price of X +$50,000). Cyberhomes and Zillow think (and report) a sale at a price of X + $50,000.

That Cyberhomes and Zillow cannot take advantage of the rich detail of data in the MLS and are only able to tap into and utilize a limited portion of the data available on each transaction in the public records is an extremely serious flaw. To me, in light of what I would call their overall "ridiculous results ratio," I'd say it's a fatal flaw.

The algorithm writers, if they wanted to get it right, would have an extraordinarily tough task that, to my mind, is somewhat analogous to the immense challenge faced by designers of language translation software.

Unfortunately, the business model of firms like Zillow and Cyberhomes is not one that requires accuracy in order to generate profits. Profits are derived from the leads they sell to realty companies and mortage companies, as well as the click-throughs to every potential profiteer in the housing market. As long as they offer speed and pizazz, there's no need to spend any R&D to take their algrithms much beyond the rudimentary and largely laughable levels that they have already achieved.

Whew!!! I feel much better now.

[sheriff]

Originally Posted by olecapt

Drop me a DM with the address and I will run you up a quick market analysis. The way the market is RE Agent have lots of time for pro bomo comps.

I am sure Eric would do the same if you prefer.

Yes, pro BOMO work from wanna-be productive real estate professionals should instill confidence in all prospective clients, and real estate professionals from other areas.

[Eric Young]

Originally Posted by sheriff

Yes, pro BOMO work from wanna-be productive real estate professionals should instill confidence in all prospective clients, and real estate professionals from other areas.

If it doesn't close, it's all pro-bono, Sheriff.

[sheriff]

Originally Posted by Eric Young

If it doesn't close, it's all pro-bono, Sheriff.

No kidding, drop me a line and I'll do a quick market analysis for you. Oh, that's right, you have a brain. Internet self-esteem, however, is a pet-peeve of mine. No offense directed to you.

[Eric Young]

Originally Posted by sheriff

No kidding, drop me a line and I'll do a quick market analysis for you.

That was the vulture swooping in.

Originally Posted by sheriff

No offense directed to you.

None taken.

[olecapt]

You are making the Sheriff's day Eric. What he loves is to get a long riff off a typo.

Best ignored.

[sheriff]

Originally Posted by olecapt

You are making the Sheriff's day Eric. What he loves is to get a long riff off a typo.

Best ignored.

Right, PKRVSOY, shoot me a DM, and I'll give you a quick analysis probomo.