NVDA intraday strategy vs buy and hold: a case study

TL;DR

The five-fold NVDA intraday selection returned 14.13% from July 2025 through September 2026, versus continuous NVDA's 49.20%, with sampled daily drawdown of 10.53% versus 21.25%. A shorter ordinary replay led 16.38% to 14.17%, but higher fees erased that limited advantage. The strategy trades NVDA and inverse NVD and closes daily. A 212-fill audit checked the three replay folds; the first native/replay discrepancy remains unresolved, and the historical development dates were repeatedly inspected.

Longer study

14.13% strategy

Matched NVDA

49.20%

Audited replay fills

212

Research completed

October 10, 2026

Follow the example

Choose a step to follow the example. Opening a request lets you review it; it does not submit a job or change your account.

Did the intraday strategy beat holding NVDA?

Over the longer study, the strategy returned 14.13% while continuous NVDA returned 49.20%. The first nine months looked more promising: independent replays returned 16.38% against NVDA's 14.17%, with less drawdown at the recorded daily sample times. Extending the same quarterly selection procedure removed that return advantage.

The strategy buys NVDA for bullish momentum and buys the inverse NVD ETF for bearish exposure. It closes positions each day. This case study documents a limited historical lead, a longer-period failure and the checks needed before treating a backtest as an edge. No live or paper trading was activated.

What the trade audit established

The three ordinary replays produced 212 filled orders and 106 closed round trips. The audit reconciled cash, quantities and fees, found no direct stock shorts or simultaneous NVDA/NVD positions, and confirmed that all audited positions closed each day. Entries met the time limits. Eighteen sampled signal events had expected trigger values and no indicator errors; the full signal-causality history was not audited.

A representative NVD sale was about 763 shares. Twelve surrounding minute bars showed 1.45 million shares traded and no zero-volume bars, but the probe returned no bid/ask quote. That checks activity near one order, not the spread or impact of every fill. The first-fold replay difference remains open. The two additional native test folds were not independently certified with an exhaustive order audit.

Exact historical measurements

The nine-month lead and its replay discrepancy

The native optimizer reported 16.20% across the first three test blocks. Separate ordinary backtests of the selected rules reported 16.38%, versus continuous NVDA's 14.17%. The last two folds matched exactly, including all 77 daily NAV samples in each. The first fold differed by 0.15821 percentage points despite the same 28 round trips. Its cause remains unresolved.

The chart below uses ordinary replay NAV. Its 230 matched observations gave 5.08% sampled drawdown versus 21.25% for NVDA. The longer chart uses the native five-fold path. Keeping the paths separate avoids silently substituting the more favorable replay result into the native study.

Higher costs erased the limited-period lead

The same selected replay portfolios were rerun at higher fees without another optimization. They retained the same round-trip counts. At 6 basis points per fill, the compounded return was 14.95%, still just above NVDA's 14.17%. At 10 basis points it fell to 9.40%.

These are actual fee reruns on submitted notional, not a deduction estimated from trade counts. Next-bar price changes are already reflected in modeled fills. A larger flat fee does not establish quote-based spread and market-impact realism. The NVDA benchmark is a continuous holding-price path without repeated strategy trading fees; this comparison does not separately certify dividend total returns.

The NVDA and inverse ETF trading rules

The bullish entry requires NVDA's five-minute return to be at least 0.75%, price above session VWAP and current minute volume at least 1.5 times the average volume of the prior 30 completed minute bars. It buys NVDA with 99% of available cash while both NVDA and NVD positions are empty.

The bearish entry requires NVDA to break below its prior 60-minute minimum (lagged one minute), trade below VWAP, satisfy the same relative-volume threshold and trade below a selected rolling mean. It buys NVD with 49.5% of cash while both positions are empty. The mean averages timestamped observed prices across a calendar-day window; it is not a conventional daily-close SMA.

NVD seeks twice the inverse of NVDA's daily return before fees and expenses. Half allocation targets roughly one-times initial bearish exposure, but does not guarantee exact minus-one-times intraday exposure. These backtests use actual NVD prices, rather than multiplying NVDA returns by a negative constant. The fund's daily objective and reset matter even when the position is held intraday.

How the walk-forward selection worked

Training expands from January 2, 2025. Each fold has an earlier training window, a five-day embargo, a validation window and then a 91-calendar-day test block. The exhaustive grid has 36 settings: rolling means of 5, 20, 50 or 100 calendar days; inverse stops of 1%, 2% or 4%; and inverse profit targets of 2%, 4% or 6%. The other entry rules, long exits, sizing and fees stay fixed. Validation percent return selects the winner before its next test block.

The five-fold continuation preserves the first three fold calendars. The first three winners use a five-calendar-day mean and 2% inverse profit target, with inverse stops of 2%, 1% and 2%. The last two use means of 5 and 50 calendar days, 4% inverse stops and 2% profit targets.

Chronological splitting limits within-study lookahead, but the researchers inspected these historical dates repeatedly while changing strategy families. All reported results are retrospective development evidence. They are not a fresh independent out-of-sample test. A future test would require freezing both the rules and the selection procedure before observing new data.

When we use the agent and when we use MCP

Codex was the research agent for this study. It proposed strategy variants, translated ideas into explicit rules, chose the next diagnostic test and reviewed the returned evidence. MCP (Model Context Protocol) connected Codex to NexusTrade's tools. The agent and MCP work together: the agent decides what to investigate; tool calls submit specific inputs to the trading engine and retrieve its results.

Use Aurora to interpret a losing strategy or inspect an assumption. A fixed replay uses the declared engine inputs; its historical result is separate from the research review.

Aurora is NexusTrade's own assistant and can help review the published evidence through the prompt on this page. MCP also exposes tools for starting a NexusTrade agent when a task needs delegated research. This study used Codex with explicit NexusTrade tool calls; it did not launch a separate NexusTrade agent. Reviewing the case study in Aurora is a new conversation, not a continuation of the historical runs.

What further tests changed

Inspect the evidence and define the next test

Private account and job identifiers are omitted. The snapshot records source-file hashes for provenance; it does not include the complete minute tape or every signal trace, so it supports inspection of the reported measurements rather than exact independent execution reproduction.

The result warrants research into when these rules work and why they miss NVDA's larger advance. It does not establish a deployable outperformer. Resolve the replay discrepancy, verify execution costs with quotes, and freeze a procedure before collecting untouched future observations. A future test can permit losing regimes while still measuring aggregate return, drawdown and costs against continuous NVDA.

Detailed measurement and provenance notes

Continue exploring