Weighted Averages

[CANative]

All of these great minds at appraisersforum and no one has discussed the use of darts. :twisted:

 

[Steve Wyrick]

I am with Mike, Richard and Nancy. In reconciliation pick your best comp, the remaining two comps provide support. Don't Average especially with 3 comps. With a large enough selection of comps then an average may have some meaning after all that is what AVM basically do. We are appraiser's not numbers crunchers like an AVM.

 

[xm39hnu]

As someone else already said, give consideration to the market in that reconciliation. Example: three equally good comps, flat or slightly declining market, go to the low end of the range. Reasoning: it likely won't sell at the high end in a dead market. Averaging in any form is defeated by market trends.

Example 2: three equally bad comps (and no others reasonably proximate); same market. Which one best represents the subject? ("Most similar" changes to "least dissmilar" in this scenario). In this situation, only one of the three is most representative of the market, so use that one's adjusted value as a baseline. Apply your market information qualitatively to this figure. Would your subject likely sell for the adjusted sales price of the "least dissimilar" comp? Higher? Lower?

Remember, all those adjusted selling prices really are is an "indication" of value, not a set-in-stone absolute. You're going for "most probable selling price," according to the definition of Market Value. Reconcile!

 

[Austin]

h34r:

 

[xm39hnu]

Austin (and I know I'm wading in dangerous waters),

How does a probability density function take into account a sluggish or very active market? I haven't seen any of those in examples where there's a time-weighted density function which would take into account a market that just changed due to external factors, i.e., new factory opening. The appraiser's qualitative analysis (which you seem to define as "divining" and "parapsycology") can account for these, and IMO such judgements (properly supported) are what we get paid for.

I like the prob-density as a way of proving most probable price, but other market factors seem to be left on the floor of the statistical cutting room. Do you know of any way to get these factors into the equations?

 

[Terrel L. Shields]

That must be the 11th standard in USPAP. Thou shalt not average.

Amen

If the available comparables are far less than ideal, why should I pick one as superior to the other? This happens with ranchland, farms, etc. If I weight all sales as equally good or bad, then 'equal weight' = sum ÷ n comparables

 

[Joker]

I use a weighted average, always have, on residential, commercial, and farm properties. I use a method similar to Bucks, but the multiplier changes on each property. The more comparable a property is, the higher the number it receives.
I use this same method in reconciling the three approches in difficult (unique) comercial appraisals also. If I have great sales data, limited but reliable cost data, and so-so income data, I would weight the 3 approaches Sales 85%, Cost 10%, Income 5%.

 

[Alan Simmons]

“Averaging” seems to be another one of those appraisal elitist areas (not that this is bad). Personally I do not see any more merit in picking the “best” comparable and using the others as mere support. Maybe someone can post an empirical study giving data to support these best practices. Short of this it is merely pomp and circumstance.

To me, the final reconciliation (within the value range) is the appraiser’s playground. The final reconciliation of value is really one of the last available opportunities for appraisers to do some professional analytical thinking. Everything else is pretty much dictated. To me, anything that locks the appraiser into a mindset (averaging or best comparable) cheapens this.

Personally I think this all stems from someone having once advised, “do not JUST average the adjusted values.” It was probably more of an attempt to get people to think than an attempt to eradicate averaging.

By the way, one pet irritant of those in the mathematical world (which I am not but I do love to torment) is the term weighted average. All averages (central tendencies) are a weight.

My reconciliation SOMETIMES reads, in part, “the above comparables give a good indication of the value of the subject and were all given weight in the final opinion of value.” To me this is as honest as I can get. I considered the three comparables and picked my final value accordingly.

 

[Bobby Bucks]

Lest we forget the Rodney Dangerfield “Back to School” real life approach....for REO
assignments use the low end of the range, for high pressure MB assignments, use the extreme high end of the range.....and if you suspect the report might be reviewed, settle for the mid range.

 

[Austin]

:lol: