One Month of Claremont Sales Is a Very Small Sample

Here is the least popular fact in local real estate reporting: Claremont is a small town, and a month is a short period, and the two together produce a sample so thin that most of the movement in it is noise.

That is not a complaint about anyone's arithmetic. The arithmetic is usually fine. The problem sits underneath the arithmetic, in the number of sales the calculation had to work with, and it does not go away because the chart looks confident.

What a small sample actually does to a figure

Every summary statistic — a median, an average, a ratio — is an attempt to describe a group with one number. How well it does that depends on how many members the group has and how similar they are to each other.

Claremont fails on both counts within a single month. The count of closed sales in a given month is small in absolute terms. And the homes are not similar: a compact condominium near the Village and a large custom house on a hillside lot are both "a Claremont sale," and they land in the same bucket.

So the summary figure is being asked to describe a handful of transactions that have very little in common. Swap two of them for two different ones and the figure moves. Nothing about the market changed. The membership of a tiny group changed.

This is the mechanism behind most of the month-to-month drama you read about small markets. It is also why what a median home price hides in a small market matters more here than it would in a city with thousands of monthly closings.

The street-level version, which is worse

Zoom in from the city to a neighborhood, and then from a neighborhood to a street, and the sample shrinks faster than most people expect. A street may go a long stretch with no sales at all. When one finally closes, it becomes the entire sample for that street.

A single sale is not a trend. It is one negotiation between two parties, shaped by that seller's timeline, that buyer's financing, the condition of that particular house, and whatever else was on the market the week the offer was written.

People do this to themselves constantly. A neighbor sells, the number travels down the block, and every homeowner within sight quietly revises their own estimate. Sometimes that revision is roughly right. Often it is not, because the neighbor's house had a renovated kitchen, or a fourth bedroom, or a lot twice the depth, or a buyer who needed to close before a school year started.

Why the conclusion is usually louder than the evidence

Reports are written to be read. A flat month produces no headline. A month that moved produces one. There is no dishonesty required for this to skew what gets emphasized — the interesting months simply get more words.

Watch for the language. "Prices climbed" and "the market cooled" are causal claims. A month of thin data can support a description of what happened to a small set of sales. It cannot, on its own, support a claim about what the market is doing.

The tell is the absence of a count. A figure published without the number of transactions behind it is asking you to trust a summary while withholding the one detail that tells you how much weight it deserves.

How to read a figure with the sample in mind

Four habits, none of which require any special access.

Look for the count first, not the figure. Before you read what the median was, find how many sales it summarizes. If the report does not say, that absence is itself information about how carefully the report was assembled.

Prefer longer windows for direction. A quarter contains more transactions than a month, and a rolling year more than a quarter. Longer windows respond more slowly, which is exactly the point: they are less easily knocked around by two unusual sales.

Treat a narrow slice as a description, not a signal. One street, one price band, one month, one property type — each cut you make shrinks the sample. Slice enough times and you are describing individual houses while using the vocabulary of a market.

Ask whether the change exceeds the ordinary wobble. Small samples move around even in a market that is doing nothing. Before you interpret a move, ask whether figures of that kind routinely bounce that much when nothing notable is happening. Frequently they do.

What a small sample cannot be rescued by

Adding decimal places does not help. Neither does a more elaborate chart, a comparison to a county figure, or a confident sentence about direction. The information content is set by the transactions that occurred; presentation cannot manufacture more of it.

Two things do help. One is a longer window. The other is more context per sale — knowing that a particular closing involved an unusual property or unusual terms tells you far more about whether to include it in your thinking than any amount of aggregation will.

THAT IS THE WHOLE TRICK. In a small market, the individual transactions are the evidence. The summary is a convenience laid over them, and its usefulness rises and falls with how many there were.

Where this leaves you if you are actually buying or selling

It leaves you needing the sales themselves, not a city-wide summary. For a specific house you want the closings that genuinely resemble it, with their condition and terms attached, and an honest account of how few of them there may be. That is a different document with a different purpose — see what a CMA answers that a market report cannot.

Anthony prepares a comparative market analysis for that work. It is not an appraisal; when a lender or a court needs an appraisal, he coordinates an independent state-licensed appraiser. The distinction matters, and so does honesty about sample size in both documents. A CMA built on three loosely comparable sales should say so rather than dress three sales up as a market.

More on how the underlying figures are constructed sits on the market reports hub, and the companion piece on the data behind Claremont market reports, and how to check it covers where the transactions come from in the first place.

Anthony Grynchal has been licensed in California since November 2009.