Is the next-day reversal weaker in shares whose daily returns tend to continue, apart from where the close sat?
Study 84 · Changed a rule or plan · recorded 4 November 2026 · answer: Partly · full paper with figures
The question
We compare the reversal, the next session's return of the day's top fifth less its bottom fifth, between shares whose daily returns over the past year had tended to continue and those whose returns had tended to reverse. Does that difference reach 0.10 percentage points on 2018 to 2022, and survive holding where the close sat in the day's range?
Data and method
Each session, shares priced at five or more in US currency at the unadjusted close are split into thirds by how closely each day's return had followed the previous day's over the past year. Inside thirds of the prior five sessions' return, the top third's reversal is set against the bottom third's and averaged over sessions, its 95% range allowing for correlation between nearby sessions. The plan rebuilt the exploratory figures and resampled 2013 to 2017 for false alarms.
The answer
Partly. Over 1,258 sessions from 2018 to 2022, among shares that traded at the next session's open and close, the next-session reversal was 0.226 percentage points smaller in shares whose returns had tended to continue than in shares whose returns had tended to reverse (95% range 0.169 to 0.283), wholly above the 0.10 set in advance. The reversal per unit of the day's move was also smaller, so the difference is not just in the size of the moves. Once where the close sat in the day's range was held in the same fit, the difference went with the close's place, and the part left with the day's own return could not be told from zero. That is the pattern a passing distortion of the closing price would leave, but daily prices cannot show it is the source.
Main estimate
0.226 percentage points (95% range 0.169 to 0.283), from 1258 observations
What it changes
Under the rule fixed in advance, a note for the record is proposed for the programme's reversal readings and the desk's reversal screens: in shares whose daily returns had tended to reverse, the larger next-session reversal goes with where the close sat in the day's range, with no source identified. The next planned reversal study holds the past-year tendency as a control beside the close's place. No trading decision changes.
What remains uncertain
These years were used by earlier studies of these shares, so this is a check inside our development data, not an independent test. Each session counts once, allowing for shares moving together on the same dates, but the side readings were each read alone: with no effect beyond the earlier years' averages, at least one of them shows a pattern in 25.5 percent of resampled runs. Daily prices cannot tell a passing distortion of the closing price from other sources.
How it was made
Run took 102 minutes; independent code review: approve; the recorded calculation reruns to the same result.
Fingerprints
plan 9fa46c614cb4ec2dfa56e9deffc25a5d81c474920b12a5b90183711e64456721 result dd62fe6d0bcb640bd3e1aa1778cf0a632c39a3bacf1b18366e9cccb3794e27ca
Exploratory work on development data; not a performance claim.