TL;DR
Buying NVIDIA calls after a 4% gap-down paid off if you held for six months. Across NVIDIA's 14 gap-downs since 2022, $7,000 of six-month calls bought at the open was down a median $311 after a month and up a median $11,066 after six months. Buying at the close made $14,480, and buying a week later made $12,637. The same calls bought on a random day lost $736.
The catch is the ride. The typical NVIDIA buy fell 58% below its high before it recovered, and across 13 chip stocks about 1 in 4 accounts that doubled down ended down half or more. On the other 12 chip stocks the stock bounced a median 19.6%, barely above the 18.9% a six-month call needed to break even, so calls made little there.
NVIDIA calls bought after a gap-down
Median across NVIDIA's 14 gap-downs. 11 of 14 were profitable at six months.
For the first month the calls went nowhere: a median -$199 a day later, -$246 after a week and -$311 after a month. Held for six months, the median was +$11,066, and 11 of the 14 buys were profitable.
Open, close, a week later, or a random day
Median across NVIDIA's gap-downs (13 have a random-day comparison).
Show the table · 14 rows, 6 columns
Scroll sideways to compare all columns. Select a heading to sort.
| 2022-05-26 | -5.5% | -$4,182 | -$5,375 | -$5,968 | +$37,015 (2022-09-27) |
| 2022-06-13 | -5.7% | -$1,687 | -$780 | -$3,815 | +$14,465 (2022-09-12) |
| 2022-08-08 | -7.8% | +$3,253 | +$3,066 | +$2,561 | -$6,577 (2022-04-25) |
| 2022-09-01 | -5.8% | +$13,096 | +$21,955 | +$12,281 | -$6,634 (2022-05-11) |
| 2022-09-13 | -4.9% | +$15,906 | +$17,731 | +$22,389 | -$1,193 (2022-07-28) |
| 2022-10-07 | -4.7% | +$24,926 | +$38,280 | +$28,586 | -$1,542 (2022-08-05) |
| 2023-10-17 | -4.5% | +$33,919 | +$33,976 | +$32,450 | +$14,644 (2023-04-06) |
| 2024-08-02 | -5.0% | -$1,134 | -$2,035 | +$3,209 | No random day |
| 2025-01-13 | -4.4% | +$6,031 | +$5,021 | +$3,075 | -$2,920 (2025-12-31) |
| 2025-01-27 | -12.5% | +$9,036 | +$11,228 | +$12,993 | +$904 (2025-01-07) |
| 2025-03-31 | -4.1% | +$25,898 | +$24,265 | +$21,502 | +$8,374 (2025-05-20) |
| 2025-04-10 | -4.3% | +$22,589 | +$22,600 | +$19,721 | -$2,247 (2025-10-10) |
| 2025-04-30 | -4.2% | +$34,566 | +$32,395 | +$18,600 | +$12,971 (2025-03-13) |
| 2025-11-25 | -4.2% | +$2,387 | +$1,609 | +$156 | -$736 (2025-10-16) |
No matching rows. Clear the filter to see all records.
NVIDIA's 14 gap-downs: six-month profit on $7,000 of calls, by entry (random day in parentheses)
Buying at the close made a median +$14,480 and buying a week later made +$12,637. Head to head, the close beat the open in only 7 of 14 gap-downs, so the higher close median is not a reliable edge.
To check whether NVIDIA's rally explains the result, I paired each gap-down with a random day. All three entries beat their random day in 11 of 13 gap-downs (p = 0.02), and the random days made a median -$736.
What I tested
Read the explanation
A gap-down is a day a stock opened 4% or more below the prior close. My options data begins in March 2022, and every test needs six months of prices after it, so the gap-downs run from May 2022 to April 2026, keeping only the first when several fell within five trading days. NVIDIA had 14, the last in November 2025. The 12 related chip stocks added 74, for 88 in all.
Each test is a $7,000 account that puts $6,930 into at-the-money NVIDIA calls expiring 190 to 260 days out, at the strike and expiry nearest those targets, with a market order. The open entry filled at 9:32 AM ET; the close and week-later entries decided at 3:55 PM and filled at 3:56 PM. NexusTrade backtests ran each one on recorded option prices.
The random day is a date, and the contract is then picked on that date by the same rule. For each gap-down I drew one trading day from the same stock and calendar year that was not a gap-down and had no gap-down in the five sessions before it. The draw is seeded, so it repeats, and it could fall before or after the gap-down (5 of 13 fell after). It is one draw per gap-down, so the -$736 median carries some luck.
If you already hold NVIDIA calls: hold or add?
Median across NVIDIA's 13 gap-downs with a prior-close entry. Each account held $7,000 of calls bought the close before the gap-down and $7,000 in cash.
Calls held through the gap-down lost a median $843 the next day, and 2 of 13 were up. Six months later, just holding made a median +$4,425.
Adding another $7,000 of calls at the open made +$13,461, and adding at the close made +$15,653. Adding at the open beat just holding in 10 of 13 gap-downs.
The catch: one real account
Portfolio value (USD)
- Account value 2025-07-30
- $16,035.99
Account value: 81.29% largest decline in plotted observations. 2025-02-20 ($9,105.99) to 2025-04-21 ($1,703.49).
On January 27, 2025, NVIDIA opened 12.5% lower. This account bought 3 calls at the open for $20.80 each. It climbed to $9,106 by February 20, 2025, then fell to $1,703 by April 21, 2025, 81% below that high at the close. By July 30, 2025 it was worth $16,036.
That was not unusual. The typical NVIDIA open buy fell 58% below its high at some point before the six months ended.
Who lost half, by stock
$14,000 accounts holding $7,000 of calls that added $7,000 more at the open, at the close, a week later, or on a second dip within the week: 288 accounts across 13 chip stocks.
Show the table · 13 rows, 4 columns
Scroll sideways to compare all columns. Select a heading to sort.
| Intel | 15 of 24 | 62% | -$9,981 |
| Arm | 11 of 29 | 38% | -$3,825 |
| AMD | 12 of 40 | 30% | +$5,515 |
| Marvell | 10 of 35 | 29% | +$5,434 |
| Qualcomm | 3 of 12 | 25% | +$477 |
| Broadcom | 2 of 8 | 25% | -$365 |
| Astera Labs | 10 of 50 | 20% | +$18,651 |
| NVIDIA | 5 of 52 | 10% | +$14,822 |
| Lattice | 0 of 16 | 0% | -$1,449 |
| MACOM | 0 of 7 | 0% | -$904 |
| Analog Devices | 0 of 8 | 0% | -$2,148 |
| Synopsys | 0 of 4 | 0% | +$3,059 |
| NXP | 0 of 3 | 0% | +$20,393 |
No matching rows. Clear the filter to see all records.
Accounts that added calls on a gap-down and ended down 50% or more, with the median six-month result
68 of the 288 accounts that doubled down ended down half or more, about 1 in 4. Intel accounts for 15 of them, but without Intel it is still 53 of 264 (20%).
NVIDIA was the exception: 1 of 13 accounts that added at the open ended down half, 1 of 13 that added at the close, and 2 of 13 that added a week later. None of the 75 accounts that just held ended down half, because half of each account stayed in cash.
Why calls barely worked on other chip stocks
Move: gap-down open to the close six months later, split-adjusted. Break-even: (strike + premium paid) / stock price at the fill, minus 1.
Show the table · 13 rows, 5 columns
Scroll sideways to compare all columns. Select a heading to sort.
| NVIDIA | 14 | +53.2% | 17.1% | Yes |
| Astera Labs | 13 | +90.9% | 26.5% | Yes |
| NXP | 1 | +35.7% | 10.3% | Yes |
| Marvell | 10 | +31.5% | 18.9% | Yes |
| MACOM | 2 | +28.6% | 17.3% | Yes |
| AMD | 10 | +25.0% | 18.2% | Yes |
| Qualcomm | 4 | +17.1% | 14.6% | Yes |
| Broadcom | 2 | +15.9% | 17.9% | No |
| Arm | 8 | +12.2% | 20.9% | No |
| Lattice | 4 | +10.2% | 17.5% | No |
| Analog Devices | 2 | +9.7% | 9.6% | Yes |
| Synopsys | 2 | +8.4% | 11.8% | No |
| Intel | 6 | -19.0% | 18.4% | No |
No matching rows. Clear the filter to see all records.
Median six-month move after a gap-down vs the median break-even of a six-month call bought at the open
The other chip stocks did bounce. Six months after a gap-down they were up a median 19.6%, and 45 of 64 were higher. A six-month at-the-money call needed a median 18.9% rise to break even, so the typical stock cleared it by a hair, and the median call bought at the open made +$1,036.
That is also why puts lost (my first study): the stocks recovered on average. NVIDIA rose a median 53.2% in the same window, about 3 times its 17.1% break-even, and all 14 were higher. 7 of the 12 other stocks cleared their break-even at the median.
What this test cannot tell you
My options data starts in 2022, so this covers four years, mostly a bull market. 72% of the gap-downs across these 13 stocks happened in 2025 and 2026. NVIDIA had only 14, which is a small sample.
A long bear market, a change in NVIDIA's leadership or new regulation after this news would fall outside what these four years show. Commissions are not included. This is research, and it is not financial advice.
Watch the video
Run the test yourself
Aurora, the AI agent in NexusTrade, can rerun this test on NVIDIA or a stock you hold. The request below uses the same gap-down rule, budget and entries, and asks for research only.
No comments yet.