Order of operations for market trend analysis

[Joe Flacco]

This is more/less how the commercial forms are laid out
View attachment 96845

It is almost exactly like top of the URAR

 

[CGinMN]

The sales comparison grid is puzzle much like sudoku or a crossword. Once you solve for one thing, other things get easier to solve, and everything is interdependent. If you start filling in wrong answers you will be led astray, so you have to start with what you know.

 

[Mejappz]

We have yet another incompetent "reviewer" interjecting themselves into the process when they don’t know what they’re talking about. A step-wise multiple regression analysis or multiple regression analysis, in general, is inherently superior in this context because it eliminates the need for a strict sequential order when determining adjustments. The sequence of adjustments becomes irrelevant when using a statistical model that simultaneously accounts for all variables and their interactions.


In a traditional paired-sales or linear adjustment process, an appraiser might first adjust for location, then for physical attributes, and finally for market conditions, assuming that each adjustment is independent. However, this manual sequencing can introduce bias and interdependencies that distort the final adjustments. Multiple regression, by contrast, controls for all relevant variables at once, meaning that the order in which adjustments are applied manually is not a concern—each variable’s influence is isolated and accounted for within the model.


In general the approach of adjusting for financing, terms of sale, and date first before physical attributes is a reasonable heuristic, but it’s an unnecessary distinction when using regression. The model will automatically allocate the correct weight to each variable without requiring a predetermined sequence, thereby reducing the potential for error caused by subjective ordering of adjustments.

 

[CGinMN]

Refining a regression through multiple iterations to add or remove variables is a form of sequencing.

 

[Mejappz]

Refining a regression through multiple iterations to add or remove variables is a form of sequencing.

Refining a regression model through multiple iterations—adding or removing variables—is a form of sequencing, but it’s not the same thing as manually applying adjustments in a predetermined order. The difference is huge.

In traditional appraisal adjustments, sequencing is rigid—adjust for market conditions first, then location, then physical attributes, etc. Each step depends on the previous one, which can introduce errors or compounding bias. With regression, the sequencing is dynamic and data-driven. The model tests variables simultaneously, measures their statistical significance, and refines itself automatically. If a variable isn’t pulling its weight, it gets dropped. If another variable interacts meaningfully with others, the model accounts for it.

So, yes, regression involves sequencing in the sense that variables are tested and refined through iterations, but it’s a completely different process from the manual sequencing an appraiser does when applying adjustments by hand. The whole point of regression is that it removes the need for an arbitrary order of operations and lets the data dictate what matters.

 

[J Grant]

Imo, the top two adjustment lines are about prices and the rest is more about value.

The first is to bring terms of sale of the prices of the comparable into equivalence per the MV definition ( a price unaffected by concessions, special financing etc)

The second is was the price affected by the sale/contract date and the market conditions at the time.

The rest relates to adjustments for property characteristics taht command a negative, positive, or neutral value - such as location, quality, view, a pool etc.

Prices are made more equivalent first, which makes sense and follows the grid; however, it can be revisited later, as can any line adjustment if indicated. Sometimes, I will add a comp or remove a comp
or replace a comp if things are out of whack after all the adjustments are made.

 

[RCA]

MARS regression first. That generates value contributions. From the value contributions, you get the adjustments. Without value contributions, you skip the constraints needed to prevent over or under-value. Yes, you can use the traditional methods and by chance get close to the value with complex properties - but you are then relying very heavily on the appraiser's subjective opinion - which makes that traditional SCA a scam, a fraud - perpetrated by the lying and idiotic leadership of the appraisal organizations. Well, excuse me, yea some of the leadership is too dumb to do anything but follow the flock so they can't be considered criminal, just dumb. But somewhere up there some time back, most likely someone who made decisions knew it was all a scam, a fraud.

Good old Sabine Hossenfelder. This is so true, even for appraisal!!

Yea, "obedient idiots" --- I could name many names.

 

[Tom4value]

I was talking to a reviewer the other day about SFR adjustments and they recounted seeing an appraiser who developed the adjustments for the location and physical attributes first before adjusting for market trends. More similar to making the apples themselves more similar prior to analyzing the seasonal variances in the market for the prices of apples. More expensive in one season than another.

I have always played around a little with the order of operations of adjusting the other variables to test the combined efficacy of different combinations but I have always adjusted for dates and financing and terms of sale first before going on to the others. I know why I did it that way but now I'm re-considering the efficacy of adjusting for the combined terms and conditions of sale prior to the combined property attributes. Maybe it makes more sense to equalize the property attributes first.

Just so happens, I took my CE course last week that addressed this. You are right. Market conditions first.

 

[Terrel L. Shields]

I would give early concessions at time of lease and then get "above market rents" with a long term lease.
Using typical cap rate for such property, I can fool the appraiser.

Not if I read the lease.

 

[TerryRohrer]

Market conditions first.

What source?