The Great Sort

The Sold Index

What things actually sold for, not what someone hoped they would.

Anyone clearing a house runs into the same two numbers: a listing price, which is a guess, and an appraisal, which is often a wish. Neither tells you what a buyer actually paid. The Sold Index is our attempt to answer that question directly — a quarterly record of realized prices for common downsizing items, compiled from public sold results on major online auction and estate-sale platforms.

Why it exists

Listing prices and appraisal fantasies mislead. Sold prices don’t.

A dining set listed at a certain price didn’t necessarily sell anywhere near it. An heirloom appraisal, done for insurance or for comfort, isn’t a market. The only number that reflects what a real buyer was willing to pay is the number a sale actually closed at. Those numbers exist, but they sit dispersed and unindexed across auction results and estate-sale archives. The Sold Index gathers them into one place, by category, on a regular schedule — so that whoever is pricing a dining set, a set of china, or a box of tools can see what similar items have actually brought at sale before anyone quotes them a number.

Planned categories

Seven categories, chosen for what actually fills a house.

Cadence

Quarterly, starting with a pilot.

The Sold Index publishes on a quarterly cadence. The Pilot Edition is targeted for the fourth quarter of 2026, covering a first pass at the categories above. Later editions will expand coverage and refine the method as we learn where the data is thin and where it’s solid.

Methodology

The rules we hold ourselves to.

  • Public data only. Every figure traces back to a sold result that is publicly visible on the platform it came from — nothing private, nothing modeled.
  • Medians, not averages. A handful of unusual sales can swing an average past anything a typical seller would see. The median holds steady against that.
  • Minimum lot counts. We require at least 25 sold lots in a category before we publish a number for it. Thin categories stay unpublished until there’s enough behind them.
  • Sources named. Every edition names the platforms the data came from, so the work can be checked, not just trusted.