Covered calls · Semiconductors

NVDA Covered Call Backtest

NVIDIA is the stock a covered-call seller most wants to own and least wants to cap. In this backtest the calls took back far more than they paid, and the gap shows up in nearly every big year.

The short answer · default setup

Monthly covered calls on NVDA trailed holding by 20.0 percentage points per year in this model: +42.8% annualized versus +62.7% for holding.

August 30, 2016 – October 15, 2026 · 0.30 delta · 30 days · buy back in-the-money calls and keep the shares. Premiums are modeled, not historical option quotes. The analysis below describes this default setup; the interactive results change with your settings.

The setup tested

On , sell a call out. If it finishes in the money, .

 

Advanced model settings

There are no live option quotes: each option is priced with Black–Scholes from the stock's own past prices (daily split- and dividend-adjusted closes from stockanalysis.com), marked up by the vol premium and trimmed by the haircut. Premiums are modeled, not market — good for comparing strategies and stocks, not for predicting exact dollars. Full methodology →

Key figures

Per year, averaged over every possible start date

Covered call vs. holding
-20.0 pts/yr
Holding+62.7%Covered call+42.8%
Finished in the money
35%
of 30-day calls finished in the money (tested from every trading day since 2016)
Income / year
24.2%
premium collected, as % of the share price, after a 5% haircut
Buyback cost / year
44.0%
paid to close calls that finished in the money; net -19.8%/yr from the calls (a simple average, so it can differ from the compounded gap vs holding)

Premiums are modeled from each stock's past prices (Black–Scholes, market-like volatility, skew, 5% haircut) — not real option quotes. How it works →

Strategy vs. just holding

Covered-call overlay Buy & hold

The verdict: NVDA covered calls trailed holding by 20.0 pts a year

I ran the standard setup on NVDA: a 0.30-delta call, 30 days out, rolled every month for 10 years. The premiums looked generous at 24.2% a year. Buying the calls back, though, cost 44.0% a year, so the net result from the options was -19.8% a year.

The position still made money: +42.8% a year. Over Oct 2021 – Oct 2026, the window I use to compare all 104 tickers, its covered-call return of +35.0% a year ranks 3 out of 104. That ranking is the trap. Plain holding returned +62.7% a year, so the covered-call version gave up 20.0 pts a year against the thing it was wrapped around.

The calls finished in the money in 35% of months, which is what a stock with a strong upward drift does to a 0.30-delta strike. Each of those months capped a stock that kept going.

May 2017, and the years that cost the most

The worst single month started in May 2017. NVDA rose 56% in those 30 days, and closing the call sold at the start of that stretch cost 50.0% of the share price. That one call cost more than the 24.2% a year the calls collected on average. A 0.30-delta strike sits, on average, 9.3% above the price. When a stock moves several multiples of that distance, the call is deep in the money by expiry and closing it takes most of the move.

The same pattern shows up in the annual numbers. In 2024 the stock returned +171.2% and the covered-call position +71.4%, a difference of -99.8 pts. In 2023 it was +239.0% against +155.0%, a difference of -84.0 pts. And in 2021, +125.5% against +55.0%, or -70.5 pts. Those were the three costliest years against holding.

YearStockCovered callDifference (pts)

Read 2020 with care: the model estimates option prices from recent price swings, which ran far above what options actually cost in the weeks after the March crash, so it overstates the premium collected then. Real buy-write funds trailed the S&P 500 in 2020.

Calendar-year total returns; the covered call is averaged over every start date, as in the chart above.

The best year against holding was 2025, at +8.7 pts. The calls came out ahead in 2018 and 2022, when the stock fell, but also in 2019 (+76.9%) and 2025 (+38.9%), so a rising year alone did not sink them. The shortfall was not simply a function of the size of the rally: 2020 was a +122.3% year for the stock and shows only -18.3 pts, but the model likely flatters that year, because option prices in the months after the 2020 crash were well below what past volatility implies. Even in the years the calls came out ahead, the edge never passed +8.7 pts. In 2022 the stock lost 50.3% and the covered-call version lost 44.4%.

The drawdown where the premium barely mattered

From Nov 2021 to Oct 2022, NVDA fell 66%. The covered-call position fell 63%. That is the stretch where the pitch is weakest for a stock like this: one month's premium is a thin cushion against a crash of this size. Typical volatility for NVDA in this data is 40% against a median of 30% across all 104 tickers, so the premium is not small. It is just small relative to what the stock can do in either direction.

How the semiconductor peers compare

Over Oct 2021 – Oct 2026, NVDA netted -20.8% a year from the calls and ranks 4 out of 8 in its sector. Closer to the top were QCOM at -11.3%, TSM at -12.1% and AVGO at -15.6%. Below it were INTC at -22.2% and AMD at -32.3%, with two more semiconductors further down. Every semiconductor I tested lost money on the calls over that window. In this sector the calls cost money across the board, and NVDA sits in the middle of a bad group rather than being an outlier.

A modeled NVDA call at the latest available close

On 2026-10-15 NVDA closed at $217. A modeled 30-day, 0.30-delta call would have a strike near $230, which is +6.0% above the price, with a premium of $2.96 a share ($296 per contract, 1.36% of the price). These are modeled premiums, not live quotes, and 100 shares cost $21,723.

The option overlays in the strategy table all lost money on net. The conservative 0.15-delta call lost the least, -11.6% a year, finishing in the money in 16% of months; the aggressive 0.40-delta call lost the most, -24.5%, in the money in 49%.

The 0.15-delta call lost the least because its strikes sat far enough out that a rising stock rarely reached them; the 0.40-delta call finished in the money almost every other month. To try other strikes and expiries on your own numbers, use the covered call calculator, and see the full rankings to compare this with other stocks.

The wheel, which sells puts to re-enter after assignment, did worse than the calls on their own: +29.6% a year, 33.2 pts a year behind holding.

NVDA wheel strategy backtest — cash-secured puts plus covered calls on the same stock.

Covered calls on similar tickers:

Compare strategies — click a row to load it

StrategyStrikeExpiry Income / yrBuyback cost / yrNet from calls / yrIn the money
Aggressive income0.40Δ30d 40.2%64.7% -24.5% 49%
Standard monthly0.30Δ30d 24.2%44.0% -19.8% 35%
45 days, 0.30Δ0.30Δ45d 19.5%38.6% -19.1% 37%
Conservative0.15Δ30d 8.8%20.4% -11.6% 16%
Far out of the money+10%30d 23.2%41.1% -17.9% 33%
Quarterly, upside-friendly+20%90d 13.0%32.4% -19.4% 38%
Net = income − buyback cost, as % of the share price. What these terms mean →

How NVDA compares: Semiconductors

Overlay helped Overlay hurt dot size = how often calls finished in the money
TickerRealized volStock ann. returnIncome / yr Buyback cost / yrNet from calls / yrIn the money
SPY S&P 500 ETF12%+13.0% 8.8%9.3% -0.5%37%
QQQ Nasdaq 100 ETF17%+14.8% 11.6%13.7% -2.0%34%
QCOM Qualcomm32%+8.9% 20.2%31.5% -11.3%25%
TSM Taiwan Semiconductor30%+31.5% 19.2%31.3% -12.1%35%
AVGO Broadcom32%+49.9% 21.2%36.8% -15.6%33%
NVDA NVIDIA40%+58.3% 25.3%46.1% -20.8%33%
INTC Intel35%+13.8% 24.5%46.7% -22.2%25%
AMD AMD48%+37.3% 26.9%59.2% -32.3%31%
MU Micron42%+70.5% 25.5%60.6% -35.1%36%
ARM Arm Holdings59%+68.0% 33.6%85.4% -51.7%33%
The same strategy run on each ticker. Above the line the calls added income; further right, the stock rose faster. Every ticker in this table is compared over the same window (named above the chart) — even where the figures at the top of the page use a longer history — so a 10-year history is never ranked against a 5-year one; faded dots have less data. Click a row to open that ticker, or see the full ranking of every ticker.
Explore the model: volatility, cycle outcomes & trade payoff

Estimated volatility & option premium over time

Estimated vol (annualized) Modeled premium ($ / contract)

Each point prices the call you'd sell that day under the current settings — volatility (left axis) comes from the stock's own recent history, and the premium it implies (right axis) rises and falls with it. This is the raw material behind every premium in the charts above.

Outcome of every cycle

Each expiry window, overlay P&L as % of stock price
Kept premium (finished OTM) Finished in the money (capped)

One trade, at a glance

Covered call Own stock only