Why can't we band together on Feb 4th?

[Non Sequitur]

It’s amazing what this profession has become. We’re trying to achieve a level of precision that is not possible.

Tell that to the puritans who believe everything we do is pure math.

 

[chad hampton]

Tell that to the puritans who believe everything we do is pure math.

Our appraisals are supposed to mirror what a typical buyer in the market would do. I’ve never seen a buyer pull out spreadsheets when they’re looking to buy a home.

But I’ve been saying for a long time, there’s a motive behind the bull**** that the suits at the national level are pushing.

 

[Overimprovement]

You all must appraise in markets where prices move in a perfectly linear direction. My market experiences significant seasonality thru the year. The shape is the same each year but the peaks and valleys change. 12 month price trend lines blanket-applied to all comps would be a joke in my market. And most markets I
suspect.

 

[Tim Hicks (Texas)]

You all must appraise in markets where prices move in a perfectly linear direction. My market experiences significant seasonality thru the year. The shape is the same each year but the peaks and valleys change. 12 month price trend lines blanket-applied to all comps would be a joke in my market. And most markets I
suspect.

We did not have seasonal markets until the past few years. Plus, sometimes a slow market does not reflect what the current market is saying. We had a notoriously slow December which showed a more downward trend than usual, but listings and pending sales are trending upward this month. So, adjust all sales this month based on a unusually low last month?

 

[Overimprovement]

We did not have seasonal markets until the past few years. Plus, sometimes a slow market does not reflect what the current market is saying. We had a notoriously slow December which showed a more downward trend than usual, but listings and pending sales are trending upward this month. So, adjust all sales this month based on a unusually low last month?

I maintain a 3 month running average table to have a more even data distribution. The variation from month to month is still not as pronounced, but even then seasonality is apparent. I've had reports where market adjustments for comp 1 are negative, comps 2 and 3 are neutral and comps 4 and 5 are positive. But I've been including my table for years. Just have to add a graph now.

 

[Non Sequitur]

I'm all for an easy button because in the end we still have to make a living despite how much money former appraisers at GSEs think we make. In any case, if you use TOTAL and are on a budget, or as many in the business now only complete a few here and there, you can buy Spark as needed at $14 for 4 uses per order (and it builds out the cost approach and extracts land value from comps too). Or you can drop your csv files into a GPT and experiment with prompts. Here's a response I got this morning, complete with fancy graphs:

"Monthly Close Price Trend Analysis:

Key Findings:

• The regression model shows an R-squared of 0.276, meaning that only 27.6% of the variation in average close prices can be explained by time.
• The trend coefficient (-$963.74 per month) suggests a slight decreasing trend in home prices over time, though it is not statistically significant (p = 0.097).
• The intercept suggests an estimated average price around $250,800 at the start of the observed period.
• The Durbin-Watson statistic (~2.68) suggests little autocorrelation in the residuals.

Interpretation:

• There is no strong evidence of a consistent price decline based on the available data.
• The scatter plot with the trend line suggests fluctuations rather than a clear upward or downward trajectory.
• External factors (seasonality, economic shifts, market demand) could explain the observed volatility."

I then asked it to analyze for seasonal trends:

"Quarterly Seasonal Analysis of Close Prices:​

Key Findings:​

1. Trend Component:
   • The red line in the second plot represents the general price trend over time.
   • The trend suggests a gradual decrease in prices, but it is not entirely linear.
2. Quarterly Seasonal Component:
   • The seasonal pattern repeats every 3 months, with fluctuations around ±3,500.
   • The values cycle between -3,539 (low), +1,383 (moderate), and +2,155 (high).
   • This suggests that prices tend to dip every third month before rising again.
3. Residual (Noise) Component:
   • The last plot shows random variations that are not explained by trend or seasonality.
   • If the residuals are high, external factors (e.g., market demand, interest rates) may be influencing prices.

Interpretation:​

• The quarterly seasonality effect suggests that home prices fluctuate within the quarter, possibly due to seasonal demand shifts (e.g., fewer closings in certain months).
• The overall trend appears slightly downward, but more data is needed to confirm this long-term."

Now getting back to if this is good enough for the GSEs, I'll say this; Most people when concluding Stable consider variations, but from what I read the puritans pushing the narrative expect a pure math answer.

 

[Non Sequitur]

My advice, take the weekend to mess around with ChatGPT or a GPT that was made for excel files (search google). Read the GSE mandate, mess around with prompts based on what they expect, and when something that reflects your market pops out save the prompts (in order) to a library so you can repeat the process. Amateur tip: There isn't any one chart or prompt that will include everything, be specific in what data you want analyzed in the csv (sale date, sale price, etc.) and how you want it analyzed (monthly, quarterly, yearly, etc.). Or just wing it and see what pops out, it's how many people learn. Save the csv file and chat to your work file. Good luck.

 

[Florida49]

Here is what I have come up with for my Chart that I will include in my reports. I know it's probably not the greatest, but I gotta start somewhere.

Market Trends Overview (Increasing)
The market area has experienced an overall upward trend over the past 12 (24) months.
Based on an analysis of market data from MLS, median sale prices in the subject’s market area have increased at an average annual rate of approximately XX per day.
The method to determine was regression analysis (See Chart). The coefficient of time in a price regression model indicated an average daily increase of $XX which is applied to the comparables (if applicable)

Pretty much the same thing (opposite obviously) will be used for declining markets.

Comments, suggestions accepted.

I'm pretty much in the dark on who to use to get a chart for the time adjustments. We don't have Spark so it has to be something else. Thanks for your help!

 

[alebrewer]

Just generated this with the following prompt:
"Generate a monthly rate of increase of 'Close Price' using 'Close Date' as the independent variable and show the monthly increase for all months."

 

[Terrel L. Shields]

My point is that one should not dismiss fireplaces as having no value

I view same as a quality issue along with rock or brick siding. At least most of my MLR that I have delved into seem to relate the two while isolating fireplace alone yielded wild differences from negative numbers to absurdly high values.