Do shares that close near the day's high earn less the next session than shares that close near the low?
Study 82 · No reading (stopped by its own checks) · recorded 3 November 2026 · answer: Not settled · full paper with figures
The question
In 2013 to 2017, among shares priced at five or more in US currency, a close at the day's high was followed by a next-session return 0.170 percentage points below one at the low, with the day's move and the prior five sessions' held. We planned to test 2018 to 2022 against a size of 0.09 points fixed in advance.
Rules fixed before the result
Fixed before any outcome: the deciding measure and its size of 0.09 percentage points; at least 1,000 sessions with 300 or more eligible shares each; and a power check that allowed no widening of the interval and needed at least 1,158 sessions. The interval was to be widened in steps until resampled earlier years raised false alarms in no more than 5 percent of runs, and any widening stopped the study.
Data and method
Each session, among shares priced at five or more in US currency at the unadjusted close with a day's high above its low, we regress the next session's return less the market fund's on where the close sat between the day's low and high, holding the day's own return and the prior five sessions' return, and average that coefficient over sessions. The plan first rebuilt this measure on 2013 to 2017 and checked its false-alarm rate by resampling those years.
The answer
Not settled: the study stopped at its planned checks, before computing any return in 2018 to 2022. The plan required that the interval around the measured gap keep false alarms at no more than 5 percent without being widened, because a widened interval would leave too little power to detect a gap of 0.09 percentage points over the 1,258 sessions. The false-alarm check, run by resampling the earlier years, called for a wider interval, so by the rule fixed in advance the test was not opened. Nothing is shown about the gap in the later years, in either direction.
What it changes
No plan changed. Under the rule fixed in advance for a stop, a corrected design is the next choice of study to make, and the two planned studies of the next-day reversal register where the close sat in the day's range as a control whose effect is still open.
What remains uncertain
Nothing after 2017 was measured, so neither a gap nor its absence is shown for 2018 to 2022. Those years were used by earlier studies of these shares, so even a measured answer would have been a check within our development data, not a test on fresh data. Daily prices also cannot tell a passing distortion of the closing price from other explanations, so a measured gap would not have named its source.