- Over 16 years, systematically selling "naked" calls on the SPX monthly expiration produced −$294,097: negative expectancy (−$1,532 per trade, Profit Factor 0.56) with a win rate of just 50.5%.
- The last equity high dates back to August 2010 (+$3,023): from there on, 15 years and 4 months of uninterrupted drawdown. And the worst losses (up to −$47,921, more than 3 times the premium) came precisely in the rebounds after crashes, when premiums look juiciest.
- The same call, sold inside a synthetic covered call, changes everything: +$309,302, 79.2% win rate, Profit Factor 1.72, 14 positive years out of 16.
- But breaking it down leg by leg reveals the truth: the synthetic underlying made +$614,508, while the short call on its own lost −$305,206 even inside the structure. The engine is the underlying, not the call.
- A covered call is equivalent to a short put: the losses cluster in crashes — −$90,203 in the 2025 tariff crash, −$89,310 during COVID, −$78,080 in the 2022 bear market. Doable, yes — but only with active risk management.
In my article on selling puts I closed with a promise: premium can be collected on both sides of the market, and sooner or later we'd take a hard look at the other side too. Here we are. Today's question is the mirror image and just as popular: is selling call options worth it? The gut answer in many textbooks is "never, the risk is unlimited." The answer from certain premium sellers is "of course, theta works for you." As always, I don't trust either camp: I backtested 16 years of real SPX options data, two complete strategies, rebuilt day by day, in dollars, with no shortcuts.
What selling a call really means
A call option is a bet on the upside: the buyer pays a premium and acquires the right to "buy" the index at a set price (the strike) before an expiration date. The seller pockets the premium up front and takes on the opposite obligation: if the index is above the strike at expiration, they pay the difference — on SPX settlement is in cash, $100 per index point.
The seller's profile is brutally asymmetric: maximum gain = the premium, not a cent more, no matter what the market does. Maximum loss = theoretically unlimited: there is no ceiling on how high an index can go. And there's an aggravating factor the put seller doesn't face: the risk is open precisely in the direction equities structurally tend to move — up.
On top of that come two practical problems. First: the margin your broker requires balloons exactly when the position goes against you, with the risk of a margin call at the worst possible moment. Second: the position is short gamma — the higher the market goes, the worse your delta gets, and losses accelerate instead of leveling off. On paper, then, selling naked calls is a terrible idea for very specific technical reasons. But I prefer to test opinions, my own included: I put the numbers to the test.
Strategy A — The naked sale: 0.5-delta calls, every month, since 2010
The strategy tested is the purest, most systematic version possible: every month, sell 1 at-the-money call (delta ≈ 0.5, strike essentially at the current price) on the standard SPX monthly expiration — the third Friday — and hold it to expiration. No stop-loss, no rolling, no discretion: entry happens right after the previous cycle expires, roughly 30 days before the next one, and the market decides the rest.
The seller collects an average of $4,804 per trade — the "full" premium of a one-month ATM option — and bets that the index won't climb above the strike within thirty days. It's the lab version of the question in the title: if selling calls has a positive expected value, this is where it has to show up. 192 trades, from January 2010 to December 2025, one for each monthly expiration.
Parameters observed in the backtest
| Underlying | $SPX (index, cash-settled) |
| Direction | Short call |
| Size | 1 fixed contract (×100) |
| Selection criterion | Delta 0.50 (at-the-money) |
| Average premium collected | $4,804 (48 pts) |
| Days to expiration | 27–32 (monthly cycle) |
| Management | None — held to expiration |
| Trades | 192 |
| Notional controlled | ~$115,000 → ~$683,000 |
Over 16 years, systematically selling naked calls produced −$294,097. Not a bump in the road: a systematic erosion. The strategy wins only 50.5% of the time, with an expectancy of −$1,532 per trade and a Profit Factor of 0.56 — and its last equity high dates back to August 2010, eight months after the test began. Unlimited upside risk isn't a footnote in the textbooks: it's a bill the market presents with regularity, because going up is exactly what equities do for a living — in our sample, 68.8% of months.
The equity line: the market's mirror image
The chart below plots the strategy's equity (red) against the SPX level (gold). You don't need a degree in statistics to see the problem: the strategy's curve is, to a good approximation, the mirror image of the index. SPX goes from ~1,150 to ~6,835 points (+494%); whoever sold calls every month paid the bill for that rally, one premium at a time. The equity's all-time high — +$3,023 — came on August 30, 2010: everything after that is downhill.
Naked call selling equity line (red) vs SPX level (gold)
Inside the 192 trades: why it doesn't work
Here the backtest debunks the most common intuition. The problem isn't losing streaks: the worst run is a mere 4 consecutive losses (−$9,766). Curiously, the longest winning streak — 7 wins, +$62,349 — coincides with the 2022 bear market. The problem is how much you win and how much you lose.
The sequence of 192 trades
Reason #1 — The payoff is lopsided. Average win +$3,827, average loss −$7,003: a payoff ratio of 0.55. With these numbers you'd need a 64.7% win rate just to break even; the observed 50.5% isn't even close. And the 6 worst trades alone add up to −$187,581 — almost two-thirds of the total loss:
| Entry | Strike | Premium | SPX at entry | SPX at expiry | SPX chg. | P&L |
|---|---|---|---|---|---|---|
| 2025-04-17 | 5305 | $15,410 | 5,283 | 5,938 | +12.4% | −$47,921 |
| 2020-03-20 | 2335 | $12,815 | 2,305 | 2,856 | +23.9% | −$39,258 |
| 2022-07-18 | 3845 | $9,910 | 3,831 | 4,258 | +11.2% | −$31,414 |
| 2024-04-19 | 4975 | $9,120 | 4,967 | 5,313 | +7.0% | −$24,695 |
| 2025-06-20 | 5980 | $11,195 | 5,968 | 6,324 | +6.0% | −$23,166 |
| 2021-03-19 | 3920 | $6,000 | 3,913 | 4,191 | +7.1% | −$21,128 |
Reason #2 — The drift works against you. In the sample, SPX at expiration was above the entry level in 68.8% of trades. When the market rose, the strategy lost an average of $4,369 per trade; when it fell, it gained an average of $4,710. But over a one-month horizon the favorable periods are structurally the minority: the put seller has the drift as an ally, the call seller has it as a permanent opponent.
Reason #3 — It's the wrong premium. The classic argument of option sellers — "implied volatility is on average richer than realized, selling premium pays" — runs into skew on the call side of the index: the market prices puts expensively (fear costs money) and calls relatively cheaply. Call sellers collect the thinnest premium in the chain while exposing themselves to the side the index drifts toward. The numbers confirm it: $371,178 in gross profits against $665,274 in gross losses. Sixteen years of premiums didn't cover even half the cost of the rallies.
📈 Performance
| Total P&L | −$294,097 |
| Gross profits | $371,178 |
| Gross losses | −$665,274 |
| Best year (2022) | $37,318 |
| Worst year (2020) | −$71,792 |
| Positive years / total | 3 / 16 |
| Profit Factor | 0.56 |
🎯 Trade statistics
| Number of trades | 192 |
| Win rate | 50.5% |
| Expectancy per trade | −$1,532 |
| Average win | $3,827 |
| Average loss | −$7,003 |
| Payoff ratio (win/loss) | 0.55 |
| Largest win | $11,614 |
| Largest loss | −$47,921 |
| Max consecutive wins | 7 |
| Max consecutive losses | 4 |
🛡️ Risk
| Max Drawdown | −$315,250 |
| Longest underwater period | 5,602 days (~15 years) |
| Time spent in drawdown | 99.6% |
| Annual P&L volatility | $25,466 |
| Sharpe ratio (on P&L, rf=0) | −0.70 |
| Sortino ratio | −0.93 |
| Net Profit / Max Drawdown | −0.93 |
⚠️ 99.6% of the time in drawdown
The most damning number in the whole backtest isn't the total loss: it's that the equity hit its all-time high — +$3,023 — on August 30, 2010, eight months after the test began, and never saw it again: 5,602 consecutive days underwater, 99.6% of the entire history, with a final drawdown of −$315,250 still open on December 31, 2025. Psychological fatigue doesn't even come into it, because there's never anything to defend: naked call selling isn't "risky": it's structurally a losing strategy.
Strategy B — The covered call: making call selling sustainable
If the problem with the naked call is unlimited upside risk, the solution is conceptually obvious: own the underlying. If the market rallies past the strike, the short call loses but the underlying gains: the unlimited loss disappears, and all that's left is giving up the upside beyond the strike. This is the covered call (or buy-write), probably the most widely used options strategy in the world.
You can't buy "the underlying" on SPX since it's an index — and in any case, at 6,000 points one contract would control a notional value of ~$600,000. That's why in the backtest the underlying is replicated synthetically with the options themselves, using put-call parity: a long call + short put at the same strike and same expiration has exactly the same P&L profile as the underlying (delta ≈ 1), for a fraction of the capital.
Each cycle in the backtest is therefore a three-leg structure, all on the same monthly expiration as Strategy A (~30 days): 1) buy an at-the-money call; 2) sell a put at the same strike (1+2 = synthetic long SPX); 3) sell a call 5 points higher — the "covered" call that collects the premium, on average $4,868 per cycle. At expiration everything settles in cash and the structure is reopened at the new prices: 192 cycles, always in the market, never managed. Here too we're measuring the "pure" strategy, on the same expirations and the same schedule as the naked sale: the only difference is the structure.
Same data, same expirations, same call sold every month — and the result isn't even in the same league: +$309,302 versus −$294,097, a 79.2% win rate, Profit Factor 1.72, 14 positive years out of 16. But before falling in love with it, the full story needs telling: by put-call parity, a covered call is mathematically equivalent to a short put. Building one doesn't eliminate the asymmetry — limited gain, potentially huge loss — it moves it from the upside to the downside, aligning it with the market's drift instead of against it. The data that follow show exactly that: lots of steady gains, and the bill arriving in the crashes.
Equity lines compared — synthetic covered call vs naked call selling (in $)
Where do the profits really come from? The leg-by-leg breakdown
This is the most instructive finding of the entire study. Breaking down the P&L of the 192 cycles into its components — the synthetic long (long call + short put) and the short OTM call — reveals a truth that almost no covered call seller knows:
Cumulative contribution of the covered call components (in $)
The weak spots: the bill arrives in the crashes
The equivalence with the short put isn't a textbook technicality: it's exactly what you see in the six worst cycles. All the serious losses come from rapid declines in the underlying, and almost all of them come from the put leg of the synthetic:
| Entry | SPX at entry | SPX at expiry | SPX chg. | Cycle P&L | of which put leg | Context |
|---|---|---|---|---|---|---|
| 2020-02-21 | 3,338 | 2,438 | −27.0% | −$85,318 | −$84,990 | COVID crash |
| 2022-04-14 | 4,393 | 3,938 | −10.4% | −$34,992 | −$34,684 | Bear market, Fed tightening |
| 2025-02-21 | 6,013 | 5,618 | −6.6% | −$31,461 | −$31,138 | Tariff crash, act I |
| 2022-08-19 | 4,228 | 3,871 | −8.4% | −$27,842 | −$27,554 | Bear market lows |
| 2025-03-21 | 5,668 | 5,300 | −6.5% | −$27,438 | −$27,140 | Tariff crash, act II |
| 2018-11-16 | 2,736 | 2,459 | −10.1% | −$22,796 | −$22,498 | December 2018 sell-off |
Drawdown (underwater) curve — the two strategies compared, in $
⚠️ The number to frame: −$90,203
Between February 21 and April 8, 2025 — 46 calendar days — the covered call lost $90,203 from peak: the equivalent of almost 5 years of average profit ($19,331/year) handed back to the market in a month and a half, with full recovery only at the end of October. It's not an isolated case: COVID cost −$89,310 in 27 days, and the 2022 bear market −$78,080, with an 812-day wait before a new high (June 2024). Overall, the account stayed below its own high 85.6% of the time. Anyone who had started at a peak, sized "for the income" rather than for the drawdown, would almost certainly have bailed at the bottom. This — not the average return — is the true measure of the risk of a passive covered call.
Year by year: two mirror-image strategies
P&L by calendar year (in $)
| Year | Naked call | Covered call |
|---|---|---|
| 2010 | −$6,347 | $7,906 |
| 2011 | −$856 | $4,149 |
| 2012 | −$14,886 | $13,809 |
| 2013 | −$24,967 | $17,606 |
| 2014 | −$13,336 | $15,137 |
| 2015 | $9,769 | $10,415 |
| 2016 | −$10,790 | $18,947 |
| 2017 | −$15,828 | $22,944 |
| 2018 | $4,280 | −$21,697 |
| 2019 | −$41,506 | $37,473 |
| 2020 | −$71,792 | $3,373 |
| 2021 | −$31,851 | $82,784 |
| 2022 | $37,318 | −$33,990 |
| 2023 | −$51,251 | $32,841 |
| 2024 | −$23,050 | $67,188 |
| 2025 | −$39,004 | $30,419 |
How much capital do you really need?
The synthetic covered call contains an at-the-money short put: the broker requires margin proportional to the notional (on average $294,346 in the sample, up to ~$683,000 in 2025). Under Reg-T, that's roughly 15–20% of notional; and margin isn't the real constraint — the drawdown is: to get through the −$90,203 of 2025 (or the −$89,310 of 2020) without a margin call, you needed a cushion far larger than the regulatory minimum.
The cost of survival
- Average annual profit: $19,331 per contract
- Historical max drawdown: −$90,203 — almost 5 average years
- Prudent capital per contract: ~$130,000–150,000
- Estimated average annual return: ~13–15% — excellent, but paid for with violent tails
- Sizing for the income rather than the drawdown means running ~4.7x leverage on the worst year
Return / risk ratios
| Net Profit / Max Drawdown | 3.43 |
| MAR (annual profit / MaxDD) | 0.21 |
| Sharpe ratio (on P&L, rf=0) | 0.52 |
| Sortino ratio | 0.69 |
| Annual P&L volatility | $36,019 |
| Time spent in drawdown | 85.6% |
So, is selling calls worth it?
✕ Naked: no, never
- Negative expectancy: −$1,532 per trade, Profit Factor 0.56
- Losses up to 3 times the premium, concentrated in post-crash rebounds
- Against the drift: SPX rises in 69% of months — and you pay every time
- Unfavorable skew: you collect the thinnest premium in the chain
- Last equity high: August 2010. Then 15 years of drawdown.
✓ Covered: yes, with judgment
- +$309,302 in 16 years, 79% winning cycles, 14 positive years out of 16
- The premium cushions sideways months and smooths the path
- Synthetic version: same exposure for a fraction of the capital
- But know what you own: it's equivalent to a short put — the risk is in the crashes
- Requires capital sized for the drawdown and regime management
The backtest's answer is clear-cut. Selling naked calls isn't "aggressive" or "for experts": it's a strategy with negative expected value, and it loses precisely because the stock market does what it has always done — go up more often than it goes down. No premium collected can make up for being systematically on the wrong side of the drift: sixteen years of premiums didn't cover even half the cost of the rallies. The covered call, on the other hand, proves that the same call can be sold sensibly: inside a structure that owns the underlying, its chronic cost becomes a cash flow that smooths returns. But let's be honest about it: the profit engine is the market, not the call, and the structure concentrates its losses in crashes. Anyone who treats it as passive income will sooner or later meet their own February 2020, or their own April 2025. Anyone who sizes it for the drawdown, understands its equivalence to a short put and manages the regime has a serious tool for monetizing time and volatility.
The variant I use in my portfolios
As with put selling, the "pure" backtest is the starting point, not the finish line. There's an additional covered call variant that I use personally and that is always present in my speculative portfolios: it targets exactly the weak spots documented above, with active risk management that adjusts exposure according to the market regime — when to be in the market, with how much exposure, how far out to sell the call — with the goal of preserving the steady income of favorable periods while cutting the tails that, in the backtest, are worth tens of thousands of dollars.
It's a version that takes study and is too complex to explain in an article — and, for now, it remains private. But if you're curious, feel free to write to me: I'm happy to reply at info@kriterionquant.com. And if you want to see how I apply the same rigor on the equity side, you can start with algorithmic trading: two different engines that complement each other.
How we ran the numbers
Methodology. Backtest on real historical options data on the S&P 500 index ($SPX), Jan 4, 2010–Dec 31, 2025, daily mark-to-market equity (5,841 observations per strategy), 1 contract per cycle, ×100 multiplier, cash settlement at expiration. Both strategies trade the standard monthly expiration (third Friday, moved up to Thursday in holiday weeks): 192 cycles each, one expiration per month, no months skipped. Strategy A: sell 1 ATM call (delta ≈ 0.5), entry right after the previous cycle's expiration (27–32 days to expiration), held to expiration. Strategy B: long ATM call + short ATM put (synthetic long) + short call 5 points OTM, same monthly expirations (entry at 27–39 days: in about a third of cycles the covered call enters 1–3 days before Strategy A — a marginal difference that doesn't affect the comparison). The last cycle of both, expiring in January 2026, is marked to market as of Dec 31, 2025. Drawdown is calculated as the distance from the prior cumulative high on the daily series; Sharpe/Sortino on daily equity changes (rf=0, annualized over 252 days). Commissions are included in the source P&L, before taxes.
This article is also available in Italian.