How ticker selection creates bias in option scans

Scans run July 7 to July 17, 2026 via Spreadsearch

Editorial note (September 25, 2026): This page was substantially revised to replace the original scan-data summary with a methodology note. The original URL and reporting period have been retained for continuity.

Purpose of the scans

The early-July tests used a curated panel rather than a random or market-wide sample. The panel was designed to make Spreadsearch testing more practical and to include different option-liquidity profiles.

This makes the results useful for product testing, but it also creates selection bias. The result set depends partly on which tickers were selected before the scan began.

Why panel construction matters

A panel made mainly of heavily traded names may return fewer wide spreads. A panel chosen because some names have previously produced wider quotes may return more. Even with identical filters, the two panels answer different questions.

The curated Spreadsearch panel was therefore not designed to estimate the average spread across the US options market. Its purpose was to test whether the tool could find and present qualifying contracts across a manageable set of names.

When should a ticker be replaced?

One empty period is not enough. A ticker can produce no qualifying result because of the active filters, the contracts available during the scan, or temporary quote conditions.

Coverage should be reviewed over several sessions. Repeated absence provides more information than one scan, although it still does not prove that every contract was tightly quoted.

The panel needs variety. It is useful to avoid a list dominated by one sector, one premium range, or one liquidity profile.

Changes reduce comparability. Replacing a ticker may improve future testing, but it also means that results before and after the change are not based on exactly the same universe.

What this means for the reports

Any observation from the panel must remain limited to the selected names and filters. A higher or lower number of results cannot be presented as a market-wide trend without a broader and more controlled sample.

For more information about the scan process and its limits, see How It Works.