Do large one-day rises with a company filing nearby keep rising from the next day's opening price?
Study 65 · Changed a rule or plan · recorded 23 October 2026 · answer: Not at the size tested · full paper with figures
The question
An earlier study found that falls with a company filing within a day kept falling from the next day's opening price while falls without one held. Whether rises do the same tells a slow response to public news, which should work both ways, from accounts that work on the downside only.
Rules fixed before the result
Fixed in advance: at least 2,000 rises of each kind and a detectable level of at most one percentage point (0.99). The plain average decided unless its range was wider than 1.4 points and the clipped average, tried on quiet shares, kept its false-alarm rate within 2.5% to 7.5% around 5%; the range was 0.66 wide, the clipped rate 5.3% (4.7 to 6.1). A range below zero but above minus one point reads as a give-back smaller than a point.
Data and method
The first test of a programme on large one-day moves took 13,963 rises from 2014 to 2022 that the market rule flags (large for the day's market, ordinary for the share's own history, mostly in the most volatile shares), 2,489 with a company filing within a day, and compared each one's return from the next opening price to the third close with quiet shares as volatile. Falls were measured alike. Ranges allow for overlapping sessions and recurring shares.
The answer
They gave back: from the next open, rises with a company filing within a day trailed quiet shares of the same volatility by 0.43 percentage points (95% range 0.10 to 0.76) on the deciding plain average. That range excludes any further rise and a give-back as large as the one-point bar set in advance; smaller give-backs fit it. Rises without a filing trailed by 0.24 (0.09 to 0.38); the gap between them on the same day and price range was -0.05 (-0.46 to 0.37). Falls flagged by the same rule with a filing fell a further 0.79 (0.43 to 1.15), 0.49 (0.06 to 0.91) behind those without, so the filing separated falls more than rises by 0.53 (-0.08 to 1.14), a range including zero. Overnight steps were small: 0.13 (-0.03 to 0.29) with a filing, 0.03 (-0.01 to 0.08) without.
Main estimate
-0.004 fraction (95% range -0.008 to -0.001), from 13963 observations
What it changes
By the rule set in advance, a give-back was a surprise: the programme wrote down what could explain it, beside the rises whose only filing came the next day (0.04 points, -0.70 to 0.77). A longer horizon was weighed and set aside as unlikely to answer clearly, so the programme asks instead whether the companies behind explained falls were doing worse beforehand. No desk rule or earlier page changes; the plan expected the rises to hold, with low confidence.
What remains uncertain
These rises were studied before, close to close, so this is a check within our development data, not a test on unseen data. A filing's content and time of day are not read, so the groups say which rises had a filing near them, not which were explained, and a filing dated the next day may have come after the opening price, which is a recorded trade, not a midpoint quote. Whether filings act more on falls is not settled.
How it was made
Run took 117 minutes; independent code review: approve; the recorded calculation reruns to the same result.
Fingerprints
plan 04fb8244f51e532e5fec171317e530b446212c8bac745ecf54cedae35d278d60 result c4a50feb5da998e41dfc90717e15ad594a238088916ca7a225cf932e0f7ece1f
Exploratory work on development data; not a performance claim.