Cross-sectional opening options pressure predicts relative 09:41-to-close returns across mega-cap names

The two pooled tests of opening options pressure asked whether the tape predicts an index proxy's own return, and both came back null. This asks a structurally different question of the same signal: on a given morning, do the names with the most buy-side options pressure outperform the names with the most sell-side pressure? The mechanism is not the same one. A cross-sectional spread differences the market component out, so a common factor that swamps a pooled time-series test cannot hide a relative effect here. A yes would have meant options flow ranks names even if it cannot time the market.

Tests the CROSS-SECTIONAL expression of opening options pressure across 7 mega-cap single names: does relative opening pressure rank relative remainder-of-day returns? A different mechanism from the pooled time-series claims (HYP-001/002, both NULL): the market component differences out of the spread. A null does not falsify index-level, trade-level, or interaction-conditioned variants. Sessions where any of the 7 names lacks a complete signal window drop whole (complete-session rule).

Seven mega-cap names over seven months of sessions. The signal is the identical opening net premium construction used in the pooled tests, frozen unchanged and computed per name: net options premium over 09:30–09:40 divided by gross premium over the same window. Each session the seven names are ranked by that signal, and the statistic is the mean return of the top two minus the mean return of the bottom two, measured as log returns from 09:41 to the close. The headline number is that spread averaged over sessions. The permutation null reassigns the signal to names within each session, which preserves that day's cross-sectional return dispersion exactly, so the test asks whether the ranking is informative rather than whether the names moved. 999 draws, preregistered positive direction, studentized session bootstrap for the interval. Sessions in which any of the seven names lacks a complete formation window drop whole.

No ranking information. The mean top-two-minus-bottom-two spread was 2.8 basis points per session with t = 0.25 and a permutation p of 0.80 over 140 complete sessions, and the rank correlation between signal and outcome was approximately zero. Differencing out the market did not uncover a relative effect that the pooled tests had masked, which is the specific alternative this design existed to test. Two things came out of it. The flow-to-direction mechanism lost its cross-sectional representation as well as its pooled one, and the lab gained its first working ranking and spread machinery, which four later cross-sectional experiments reused frozen rather than rebuilt.

Formation windows close roughly a minute before outcome windows open, and measured feed publication latency consumes most of that slack. Live availability of each signal at the formation boundary is therefore tight rather than generous.

A top-two-minus-bottom-two spread over seven names is a blunt instrument: it uses the extremes of a narrow cross-section and discards the middle, so it has less power than a ranking test over a broad universe would. The null is evidence about this signal in this population, not about cross-sectional flow effects at large.

The registered claim covers the aggregate representation of opening pressure. A null does not falsify index-level, trade-level, or interaction-conditioned variants, and the lab registered those it could instrument as separate experiments. Like every result the lab has published so far, it is a discovery-substrate result: preregistered, but not yet confirmed on data the research process has never seen.