Nvidia (NVDA): Persistent strength with calmer recent trading

insufficient. Registered 2026-10-01.

Too few non-overlapping observations in one or more periods. The result did not improve on the baseline after the cost allowance in every period. The annual comparison did not pass the 512-trial correction (adjusted p=1.0000).

Higher 182-calendar-day company returns minus the S&P 500 price return on identical dates; 0.4 percentage-point round-trip cost allowance. Non-overlapping windows; period-end windows excluded. Baseline: eligible dates in the same period, also measured relative to the market. Positive median market-relative returns after costs are also required in all three periods. Annual sign comparisons use a fixed 512-trial Bonferroni correction. Discovery before 2010, validation 2010–2017, confirmation 2018 through September 2026. Conditional association, not a risk-adjusted alpha estimate or an executable trading strategy. Historical selection remains exploratory; forward observations start after publication.

Open the evidence, sources and forward ledger

Conditions

• $86580.ret_365d:gt:20
• $86580.sma200_gap:gt:0
• $86580.vol21:lt:35

Research history

• 2026-10-01T03:19:51.660Z — The conditions match the available data for 2026-09-30.
• 2026-10-01T03:19:51.625Z — Too few non-overlapping observations in one or more periods.
• 2026-10-01T03:19:39.538Z — Investigation registered. Its conditions and test periods are fixed.