From the next opening price, do shares that fell with no company filing still beat those with one?
Study 61 · Changed a rule or plan · recorded 20 October 2026 · answer: Yes · full paper with figures
The question
An earlier study found falls with no company filing within a day recovered more over three sessions than falls with one, two thirds of it on the first session, where the gap between buying and selling prices can make a rebound. We measure from the next opening price, leaving out the first session's step from the fall's close to it.
Rules fixed before the result
Fixed before any outcome: at least 2,000 falls of each kind, and a smallest detectable difference of at most one percentage point; it was 0.69. The plain average decided unless its 95% range exceeded 1.4 points; the clipped average could then decide only if its false-alarm rate on the same companies' ordinary spans lay within 2.5% to 7.5% around a 5% target. The range was 1.2 points, so the plain average decided; the clipped rate was 5.3% (4.6 to 6.0).
Data and method
We took the earlier study's 16,957 falls by 2,089 companies from 2014 to 2022, all with a next-day opening price, and compared, within each day and price range, each fall's return from that price to the third session's close, less the market's, for falls with and without a company filing within a day. A filing's content and time of day are not read, so the groups say which falls had a filing near them, not which were explained.
The answer
Yes. From the next opening price to the third session's close, falls with no company filing beat falls with one on the same day in the same price range by 1.5 percentage points (95% range 0.8 to 2.1) on the plain average, which decided, and by 1.2 (0.7 to 1.6) on the clipped average. Against what these shares do over ordinary three-session spans, falls with a filing lost a further 1.4% (1.1 to 1.8) while those without were flat (-0.2%, -0.7 to 0.2). The overnight step's own gap was 0.6 points (0.1 to 1.0), 30% of the gap from close to close on the same falls. Descriptively, the gap was 2.4 points (1.0 to 3.8) over sessions 1 to 5 and 0.3 (-0.7 to 1.4) over sessions 6 to 20, and the no-filing falls' return was not graded by volume.
Main estimate
0.015 fraction (95% range 0.008 to 0.021), from 16957 observations
What remains uncertain
These are the earlier study's own falls, already read from the close, so this is a check within our development data, not a new test. The opening price is a recorded trade, not the midpoint of buying and selling prices, so part of a rebound can remain: without the first session the gap was 1.2 points (-0.1 to 2.5), not excluding zero. Filings are not news, and the ranges group falls by day but not neighbouring days' overlapping sessions.
How it was made
Run took 127 minutes; independent code review: approve; the recorded calculation reruns to the same result.
Fingerprints
plan cd25c5cf0cc23caa47678e25932db5d19a9e914abedc5482d81cee385c779ea5 result d2fd8f62cfa4e9126193d27aa22d72f579179242604a3eab8d20db3e308c2ff3
Exploratory work on development data; not a performance claim.