Defined risk vs naked income: does the protective wing pay off?

Sell a cash-secured put and a crash can hand you a brutal month - if the stock gaps down, you're on the hook all the way to zero. The fix the textbooks sell is defined risk: buy a cheap, far-out-of-the-money option as a wing so your maximum loss is capped. But that wing isn't free - you pay for it every single month, and most months it expires worthless. So does the protection pay for itself, or just bleed? We backtested it 21 years on the S&P 500.

The clean answer: the wing caps the disaster - that part is real and dramatic - but it is insurance, and like insurance it normally costs a small premium rather than paying its own way. The popular "spreads beat naked puts" backtests that show otherwise are usually pricing the wing too cheaply.

The scorecard (2005–2026)

Three put-selling strategies, same 30-delta short put (same exposure), all cash-secured, held to expiration. The only difference is what protection you bolt on:

Look at the worst-month and drawdown columns first - that's where the story is. The naked put's worst month was -28.9% (the March 2020 crash); the bull put spread's was -5.3%. The drawdown fell from -28.8% to -10.6% - roughly -18 points shallower - at less than half the volatility. The wing did its one job: it capped the catastrophe.

Growth of $100,000, 2005–2026

Here is where you have to read carefully. In the model the defined-risk lines actually finish above the naked put - but don't take that at face value. That outcome leans on the wing being priced too cheaply (more on that next). What the chart shows reliably is the smoothness: the spread and condor barely flinch in 2008, 2020 and 2022, while the naked put lurches.

Did the wing pay for itself? In the model, almost - and that's the catch

Over 256 months, the protective put wing cost about 71% of collateral in cumulative premium and paid back about 74% in crashes - a 1.05x payoff-to-cost ratio, essentially break-even. Taken literally, that says the crash protection was nearly free.

It wasn't. We price every option off VIX, which is an at-the-money volatility - and far-OTM index puts trade at much higher implied vol than that (the volatility skew is steepest exactly where the wing lives). So the modeled wing is meaningfully cheaper than the one you'd actually buy. We stress-tested this with a realistic skew (a 10%-OTM index put commonly carries 5-10 extra vol points): the payoff-to-cost ratio falls from 1.05 to roughly 0.4-0.8 - the wing returns 40 to 80 cents on the dollar, a real and persistent drag. That is the normal, expected state of affairs: insurance that caps your tail usually costs a premium.

We stopped estimating: the real-fill answer

We've since bought real OptionsDX end-of-day SPY chains and re-ran all three strategies on traded bid/ask prices, 2010–2023 (163 cycles), strikes chosen by real delta. The stress-test prediction lands almost on the nose: the wing's real payoff-to-cost ratio is 0.64x - inside the 0.4-0.8 band we projected, and well under the model's 1.05x break-even. The wing is a real, persistent drag, exactly as insurance should be.

Real prices flip the ordering to the honest one. The model had the defined-risk lines finishing above the naked put; with real fills the naked put leads on return (3.51%), the bull put spread trails by about a point (2.45%), and the iron condor goes slightly negative (-0.34%) - its short call got run over again and again in the 2010s bull market, exactly the "optimistic ceiling" the model's caveats flagged. You pay for protection; you are not paid to hold it.

But the reason to buy the wing survives completely intact, because it never depended on pricing: the bull put spread's worst month was -6.3% against the naked put's -28.5% (the March-2020 crash), and its max drawdown -9.0% against -29.0%. Skew makes the hedge cost more, not work less. Trust the risk numbers; the return edge the model gave the spreads was the illusion.

The iron condor: income from both sides, capped both ways

The iron condor bolts a bear call spread onto the bull put spread, so you collect premium from a short put and a short call, with a wing on each side. In the model it earned the most (6.36% a year) at a similar capped drawdown (-11.4%) - both tails defined. Two honest caveats: it now loses when the market rips through your short call (you've capped the upside you were implicitly long), and its modeled income is the most skew-distorted of the three, because VIX overprices the short call while underpricing both wings. On real fills its return doesn't just slide - it collapses from the modeled 6.36% to -0.34% (the real-fill table above), slightly negative, the worst-distorted of the three. Read its modeled return as a fantasy ceiling; the durable feature is that no single month - crash or melt-up - can blow it up.

What this means for how you trade

• Defined risk is insurance, priced like insurance. The wing caps your worst case (-29% → -5% in the crash month) for a small recurring cost. You're paying to make the disaster survivable, not to boost returns.
• It also frees capital. A spread's margin is its width, not the full strike - so the same account can hold far more of them. That's real efficiency, but it's also how defined-risk traders quietly add leverage; size to the max loss, not the count.
• Pick naked if you have the capital and the stomach to hold a -29% month, and want to keep every cent of premium in calm years.
• Pick defined risk if a single gap-down month could blow up your account or your nerve. The cap is cheap relative to what it prevents - and you can model the exact trade in the bull put spread calculator or the iron condor calculator.

Caveats - read these

• The return comparison is flattered by skew; the risk comparison is not. Premiums are modeled from VIX, which underprices the far-OTM wings (and overprices the iron condor's short call). So the defined-risk returns here are optimistic - realistically the wing is a net cost. The drawdown and worst-month figures are structural and robust.
• Same notional, cash-secured, no leverage. All three are sized to the same 30-delta short put and fully collateralized, to isolate the effect of the wing. A spread run on its (much smaller) true margin would scale both the return and the risk up.
• Idle cash earns 0%, no commissions or slippage, held to expiration with no management - a mechanical baseline. Real spreads carry four legs of bid-ask, which hurts defined risk more than naked.
• One window, one underlying. SPY, 2005-2026, 30-delta shorts and 10-delta wings. Different strikes or a single stock would shift the numbers. Past performance is not predictive; educational, not advice.

Method: real daily SPY and VIX, 2005-02-18 to 2026-05-15 (256 monthly third-Friday cycles). 30-delta short puts/calls and ~10-delta wings, priced with the same Black-Scholes engine as this site's calculators; all three strategies cash-secured to the short put strike (same notional). The "real-fill" section re-runs all three on real OptionsDX end-of-day SPY chains (2010–2023), every leg at the bid/ask mid, strikes by actual delta - four independent backtests agree, and the naked leg reproduces the cash-secured-put ladder's 3.51% exactly. Every figure regenerates from the underlying data; none are hand-entered.