Nasdaq-100 (Oct 2023), overnight-gap reversal

The rules

At each session's first minute, test whether the overnight gap from the previous official close is a jump: the gap joins the previous session's one-minute returns, and Barndorff-Nielsen and Shephard's ratio test of realized against bipower variation, scaled by the tripower quarticity, must exceed 3.09, with the gap also the largest move in the sample. A jump down is bought and a jump up sold short at 09:31, 1% of the account each, and the position is closed at 11:30. A margin account, because it sells short.

Universe

The 93 Nasdaq-100 members on October 27, 2023, the last session before the test, that still traded under their tickers on October 30, 2026: AAPL, ABNB, ADBE, ADI, ADP, ADSK, AEP, ALGN, AMAT, AMD, AMGN, AMZN, ASML, AVGO, BIIB, BKNG, BKR, CDNS, CEG, CHTR, CMCSA, COST, CPRT, CRWD, CSCO, CSGP, CSX, CTAS, CTSH, DDOG, DLTR, DXCM, EBAY, ENPH, EXC, FANG, FAST, FTNT, GEHC, GFS, GILD, GOOGL, HON, IDXX, ILMN, INTC, INTU, ISRG, JD, KDP, KHC, KLAC, LCID, LRCX, LULU, MAR, MCHP, MDLZ, MELI, META, MNST, MRNA, MRVL, MSFT, MU, NFLX, NVDA, NXPI, ODFL, ON, ORLY, PANW, PAYX, PCAR, PDD, PEP, PYPL, QCOM, REGN, ROST, SBUX, SNPS, TEAM, TMUS, TSLA, TTD, TXN, VRSK, VRTX, WDAY, XEL, ZM, ZS

Bars

1-minute

Position size

1% of equity per stock at entry, never rebalanced

Rebalance

Every session: enters at 09:31 on a jump, closes at 11:30

Results

Hypothetical: a backtest with modeled costs, not traded with real money.

Hypothetical growth of $10,000

Rebuild it in Orbital

Paste this into Orbital Chat. The Strategy Developer builds the workflow and asks before it runs any backtest.

Trade AAPL, ABNB, ADBE, ADI, ADP, ADSK, AEP, ALGN, AMAT, AMD, AMGN, AMZN, ASML, AVGO, BIIB, BKNG, BKR, CDNS, CEG, CHTR, CMCSA, COST, CPRT, CRWD, CSCO, CSGP, CSX, CTAS, CTSH, DDOG, DLTR, DXCM, EBAY, ENPH, EXC, FANG, FAST, FTNT, GEHC, GFS, GILD, GOOGL, HON, IDXX, ILMN, INTC, INTU, ISRG, JD, KDP, KHC, KLAC, LCID, LRCX, LULU, MAR, MCHP, MDLZ, MELI, META, MNST, MRNA, MRVL, MSFT, MU, NFLX, NVDA, NXPI, ODFL, ON, ORLY, PANW, PAYX, PCAR, PDD, PEP, PYPL, QCOM, REGN, ROST, SBUX, SNPS, TEAM, TMUS, TSLA, TTD, TXN, VRSK, VRTX, WDAY, XEL, ZM and ZS on 1-minute bars, built as a program. At 09:30 each session, add the gap from the previous official close to the previous session's one-minute returns and run Barndorff-Nielsen and Shephard's ratio jump test (realized against bipower variation, scaled by tripower quarticity). If the statistic exceeds 3.09 and the gap is the largest move, buy a gap down or sell short a gap up with 1% of equity, and close the position at 11:30. Use a margin account. Backtest from October 30, 2023 to October 30, 2026 with $1,000,000.

Is it real?

Strategies tested together

15

Beta to SPY

0.01

Alpha vs SPY

−10.7% a year

Costs paid

$230,117, 416% of the gain before costs

Data

1m decisions on 1-minute bars, 754 sessions

Fills

Next open at market, sized at the fill

Read this before you trust it

This basket uses no hindsight: the Nasdaq-100 as it stood the day before the test began, from its published membership. Alphabet is held once, as GOOGL, and seven members are out because they stopped trading here before the test ended: SGEN, SIRI, ANSS, WBA, AZN, EA and WBD. Its runs start with $1,000,000 so that a 1% share buys whole shares of the dearest members; at $100,000 a stock above $1,000 could never be bought. Picked on November 8, 2026 by a rule written before any run: published evidence, yearly costs at most a quarter of the published gross return, buildable in Orbital, and the source's own parameters. Every rule is published whatever it shows. Orbital's engine cannot rank entries for a program that decides more often than daily, so instead of the source's position cap each stock gets an equal share and every signal is taken. The selection expected it to lose money: it trades a few times a day, and its published edge is fastest at the open.