Do the day's biggest risers closing at their highest price of the past year rank higher on the next day's return, chiefly overnight?
Study 87 · Changed a rule or plan · recorded 5 November 2026 · answer: Partly · full paper with figures
The question
Risers are each day's top fifth by return among shares priced five or more, ranked 0 to 1 on the next day's return. Is the lead of those closing at the past year's high over other risers above zero across its 95% range, its top at least 0.010 (one percentile point), and larger overnight than in next-day trading?
Rules fixed before the result
Fixed in advance: yes if the 95% range lies above zero and its top reaches 0.010 (one percentile point), more of the advantage sits overnight than in next-day trading by a reading passing its chance check on earlier years, and finer volume and missing-day checks do not undo it. Otherwise a range above zero, partly; one wholly below zero, not settled; any other topping under 0.010, not at the size tested; else, or failed count or power checks, not settled.
Data and method
Each session, the risers' next-day returns, less the market fund's, are ranked after a straight-line adjustment for the day's own gain and for where the close sat in the day's range, compared only among risers with similar trading volume. The mean rank of those that closed at their highest price of the past year is set against the others', with a 95% range allowing for correlated nearby sessions. Resampled 2013 to 2017 sessions set each reading's chance rate.
The answer
Partly. Over 801 sessions from 2018 to 2022, among the day's biggest risers among shares priced five or more with a usable next day of prices, those that closed at their highest price of the past year ranked 0.018 higher on the next day's return than the other risers (on a scale of 0 to 1, where 0.010 is one percentile point; 95% range 0.012 to 0.023), so the whole range lies above the 0.010 set in advance. The rank advantage was larger in the overnight gap than in the next day's trading hours, but on earlier years a pattern like that arose by chance too often to count, so which part carries it is not settled. The advantage remained when each share's trading volume relative to its usual level was also adjusted for.
Main estimate
0.018 fraction (95% range 0.012 to 0.023), from 801 observations
What it changes
Under the rule fixed in advance, a close at the year's highest price is recorded as marking risers that rank higher the next day, with the part of the day carrying it left open; no note for the record is proposed until that part is read. The next planned test crosses this condition with whether a company filing explains the rise. No trading decision changes.
What remains uncertain
Earlier studies used these years, so this is a check inside our development data, not an independent test. The adjustment is a straight line within volume bands and leaves out risers without a usable next day; finer checks bound what that leaves. Each session counts once, allowing for shares moving together on a date; side readings were each read alone, and with no effect beyond earlier years' averages at least one shows a pattern in 24.6 percent of resampled runs.
How it was made
Run took 94 minutes; independent code review: approve; the recorded calculation reruns to the same result.
Fingerprints
plan 8fd84d67738f7d4f148e75d6ac87f11db3644fb48386f1c7db42dcc40faf0a1c result c9e80bcb68567301ef28f947491cb70439d1581cbd26665229f1171c2e28d612
Exploratory work on development data; not a performance claim.