Does the gap in five-session returns after reverse splits also appear when ordinary sessions replace the splits?
Study 3 · No reading (stopped by its own checks) · recorded 26 September 2026 · full paper with figures
Correction, 30 September 2026. Technical wording rewritten in shorter sentences under the readability rule; the meaning is unchanged.
Correction, 19 October 2026. This text was reviewed after the study finished, in a catch-up of the earliest studies; the result, outcome and dates are unchanged.
The question
In an earlier study, five-session returns after a reverse split, which merges a company's shares into fewer, were a median 3.87 percentage points below the company's average over its ordinary sessions. We put an ordinary session of the same company, drawn at random, in place of each split session, to see whether the comparison alone gives a gap that size.
Rules fixed before the result
The rules for each possible answer were written down and fixed before any result was seen.
Data and method
No split session's own return was used. In each of 10,000 repetitions we put a randomly chosen ordinary session of the same company in place of each split session and measured the same median gap against that company's other ordinary sessions, which shows how large a gap the comparison produces with no split involved. Before any reading, a built-in check on the test itself had to pass.
The answer
No answer: a built-in check on the test itself failed, so by the rule set in advance no result was used.
What it changes
Nothing changed, because no result was used.
What remains uncertain
Whether the gap is real is still unknown. These are the same reverse splits the earlier study examined, and the test assumes a split session is interchangeable with the company's ordinary sessions at least 63 sessions from any split, which may not hold for the splits concentrated in 2017 to 2019. Only the five-session window was studied.