RCA
Elite Member
- Joined
- Jun 27, 2017
- Professional Status
- Certified General Appraiser
- State
- California
Oh My God! Here comes another class on adjustments:
mailchi.mp
georgedell.com
George warns in his latest email from today:
By George Dell on December 18, 2024[/td]
"...Time-series analysis is an established method in economic measurement."
You don't say?
1. Yes, time-series analysis is an established method of analysis ( it is not measurement. I get the impression George has read Douglas Hubbards "How to Measure Anything: Finding the Value of Intangibles in Business" - which loosely states that if we define a sufficiently large degree of uncertainty for what we call a measurement, then an estimate can be called a measurement. However, to say time-series analysis is an established method of economic measurement is not true, although it is an established method of economic analysis. To refer to estimates as measures, you had best reference Hubbard's quantifiers, i.e. degree of uncertainty, and oh well that is indeed nearly impossible to do with date of sale against a background of many other price factors.
2. To be clear it is not suitable for finding adjustments for residential sales in many of the metro areas in California. The day of "seasonal" changes no longer exists, IMO. We live in a more complex world.
3. - I'm pretty sure they ask ChatGPT or some other AI if they are on the right track. Only I would add here, you have to be smarter than AI to get the right answers. And they are not. I typically run into wrong answers from ChatGPT, Claude and Grok-2. I challenge them with my corrections and they invariably change their minds.
4. Time series can be layered in with other approaches to extract seasonal components. And, I have seen significant seasonal components at times in the past. .... Prices would used to go down in the winter months and up in the summer, now prices often just don't go down in the Winter, and in t. We still see that to a much smaller extent in some areas, in some years. But it can be easily disrupted by other more important events, events which are often unpredictable. Pandemics, war, elections, fed decisions, ..... But that is NOT the killer for time series, it is something else.
Here is Burlingame, CA for the past 5 years, and see if you can find any consistent patterns.
Subscribe to the Analogue Blog
mailchi.mp
Home - George Dell, SRA, MAI, ASA, CRE %
George Dell, SRA, MAI, ASA, CRE. Contact Us Analogue Blog by George Dell. Free Weekly Blog Posts. Stats, Graphs, and Data Science Classes
georgedell.com
George warns in his latest email from today:
By George Dell on December 18, 2024[/td]
"...Time-series analysis is an established method in economic measurement."
- "Most of the advice given today is questionable. At best, it is vague and confusing. It lacks step-by-step instructions on a reliable, repeatable method which meets the requirements below:"
You don't say?
1. Yes, time-series analysis is an established method of analysis ( it is not measurement. I get the impression George has read Douglas Hubbards "How to Measure Anything: Finding the Value of Intangibles in Business" - which loosely states that if we define a sufficiently large degree of uncertainty for what we call a measurement, then an estimate can be called a measurement. However, to say time-series analysis is an established method of economic measurement is not true, although it is an established method of economic analysis. To refer to estimates as measures, you had best reference Hubbard's quantifiers, i.e. degree of uncertainty, and oh well that is indeed nearly impossible to do with date of sale against a background of many other price factors.
2. To be clear it is not suitable for finding adjustments for residential sales in many of the metro areas in California. The day of "seasonal" changes no longer exists, IMO. We live in a more complex world.
3. - I'm pretty sure they ask ChatGPT or some other AI if they are on the right track. Only I would add here, you have to be smarter than AI to get the right answers. And they are not. I typically run into wrong answers from ChatGPT, Claude and Grok-2. I challenge them with my corrections and they invariably change their minds.
4. Time series can be layered in with other approaches to extract seasonal components. And, I have seen significant seasonal components at times in the past. .... Prices would used to go down in the winter months and up in the summer, now prices often just don't go down in the Winter, and in t. We still see that to a much smaller extent in some areas, in some years. But it can be easily disrupted by other more important events, events which are often unpredictable. Pandemics, war, elections, fed decisions, ..... But that is NOT the killer for time series, it is something else.
Here is Burlingame, CA for the past 5 years, and see if you can find any consistent patterns.
Last edited: