# Options Data Studies: Real Backtests, Real Numbers Source: https://theoptionsbench.com/data-studies/ > Plain-text mirror for AI/LLM ingestion. Canonical HTML page: https://theoptionsbench.com/data-studies/ --- Updated 3 September 2026 ยท by Theo Chen Options income is sold with cherry-picked numbers and round-number yields. We'd rather show the data. Below are 9 backtests on **real market history** - real prices, real index series, real option chains - and the short answer they share is uncomfortable: **no options-income strategy we tested beat simply holding the index on total return.** What selling options actually bought was a smoother ride - lower drawdowns, not higher returns - and the strategies that won the most often tended to make the least. Every figure below regenerates from the underlying data, and the correctness-critical ones are independently re-derived. Every dataset is also published as raw JSON and CSV, licensed CC BY 4.0 - cite the numbers directly instead of re-typing them off a page. ## What the data shows - No income strategy beat buy-and-hold on return: Cash-Secured Puts 6.9%/yr, the Wheel 6.5%/yr and Short Strangles 0.8%/yr all trailed the index's 10.8%-12.7%. - A high win rate is not safety: the far-OTM Put won 95% of months and the 5-delta Short Strangle 96% - yet both still took a severe crash loss. - The real, repeatable benefit of selling options was lower drawdown: put-writing cut the S&P's worst fall from 55% to 37%. - A distribution yield is not a return: QYLD paid ~14% a year yet its price fell 28%, returning +172% while the Nasdaq-100 returned +840%. - Defined risk is insurance priced like insurance: on real fills the Bull Put Spread's worst month (-6%) ran far shallower than the naked Put's (-29%), for a small, persistent cost. ## How the strategies compare Every study at a glance - the strategy, the data behind it, the headline result, and the verdict: | Strategy | Data & window | Result | Verdict | | Cash-Secured Puts | CBOE PutWrite index, ~19 yrs | 6.9%/yr vs S&P 10.8%; drawdown 37% vs 55% | Lower return, far lower risk | | Far-OTM Puts | SPY 10-delta CSP, 2010-2023 | 95% of months won, 1.3%/yr; calendar 2020 -19.8%, worst month -25.1% | High win rate, no cushion | | The Wheel | SPY real fills, 2010-2023 | 6.5%/yr vs 12.7%; drawdown 29% | 51% of holding's annual return | | Covered-Call ETF (QYLD) | QYLD vs Nasdaq-100, ~12.5 yrs | +172% vs +840%; price -28% at ~14% yield | A yield is not a return | | Selling when IV is high | SPY real fills, 2010-2023 | Sell-always 6.0%/yr beat timing 4.4%/yr | Timing lost to always-on | | Expected-move accuracy | SPY vs VIX, 21 yrs | Inside 1-SD 83.3%, 2-SD 98.6% | Beats textbook; tails violent | | Defined risk vs naked | SPY spreads, real fills, 2010-2023 | Worst month: naked -29%, spread -6% (separate sell-offs); drawdown 29% vs 9% | The wing caps the disaster | | Short Strangle (7 strike deltas tested) | SPY real chains, 2010-2023 | 78% won, 0.8%/yr vs price-return buy-and-hold 10.7%; no best strike | High win rate is bait | | Managing trades (50% / 21 DTE) | SPY daily marks, 2010-2023 | 21-DTE exit beats holding 26.8% of months; whole edge is March 2020 | Insurance, not free money | Returns are annualized (CAGR); "drawdown" is the worst peak-to-trough fall. Windows differ by study - each page states its own. Buy-and-hold is total return where a real series exists, price return where noted. ## The studies in depth ### Cash-secured puts vs buy-and-hold The real CBOE PutWrite index returned 6.9% a year to the S&P's 10.8% over ~19 years - lower return, but it cut the worst drawdown from 55% to 37%. Read the study -> Data: JSON CSV 6.9% vs 10.8% put-write vs buy-and-hold ### Are far-OTM puts "safe" income? (a myth) The far-out-of-the-money 10-delta Cash-Secured Put won 95% of months but made just 1.3% a year. It returned -19.8% across calendar 2020; its worst month was -25.1%, versus -29.6% at-the-money. Thin premium is no cushion. Read the study -> Data: JSON CSV 1.3% / 95% win 10-delta return vs win rate ### The Wheel strategy, backtested Sell Puts, get assigned, sell Calls, repeat. Over 2010-2023 (14 years) of real OptionsDX fills the full Wheel earned 6.5% a year to buy-and-hold's 12.7% - at a 29% maximum drawdown, 22 assignment cycles, 25% of the time holding shares. Read the study -> Data: JSON CSV 6.5% vs 12.7% Wheel vs buy-and-hold ### Does QYLD's 14% yield actually pay you? QYLD's distribution returned +172% since 2013 while the Nasdaq returned +840% - and its share price fell 28%. A distribution rate is not a return. Read the study -> Data: JSON CSV 28% price drop while paying ~14% yield ### Does selling options when IV is high work? We tested the "sell premium when IV is rich" maxim on 2010-2023 of real OptionsDX fills. Timing by IV rank made each trade richer and the ride smoother, but it fired only 11% of the time, so selling every month (6.0%) beat it (4.4%) on total return. Read the study -> Data: JSON CSV 6.0% vs 4.4% sell always vs time it ### Does the stock stay inside the expected move? Over 21 years, SPY stayed inside the VIX 1-SD expected move 83.3% of the time and the 2-SD band 98.6% - more than the textbook 68%/95%, because returns are fat-tailed. But the rare breaches were violent: Feb-Mar 2020 was a 7.9-sigma move past the band. Read the study -> Data: JSON CSV 83.3% / 98.6% inside 1-SD / 2-SD band ### Defined risk vs naked: does the wing pay off? On real OptionsDX fills, the naked Put's own worst month ran -29%; the Bull Put Spread's worst month over the same window, a separate sell-off, was just -6% - and its drawdown ran 9% against the naked Put's 29%. Defined risk is insurance - it caps the disaster for a small premium. Read the study -> Data: JSON CSV 29% to 9% drawdown, naked to spread ### The short strangle: a high win rate that still loses Sell an OTM Put and Call, win most months - and still lose. A monthly SPY Short Strangle won 78% of months yet made 0.8% a year to buy-and-hold SPY's price return of 10.7% (dividends excluded), because one 2020 crash erased 42 winners - and no strike was measurably better. Read the study -> Data: JSON CSV 0.8% vs 10.7% strangle vs buy-and-hold (price) ### Does managing trades (50% profit, 21 DTE) work? The 50% profit target is not crash insurance: its worst month matched holding's in 4 of 5 strategies. The Cash-Secured Put was the exception, improving from -$8,716 to -$3,860, a 55.7% reduction after the target closed a winner before a later crash. That is sequence luck, not protection for a losing trade. Read the study -> Data: JSON CSV + JSON CSV 4 of 5 profit-target worst = hold worst ## How we run these - and why you can trust the numbers Where a real published series exists we use it, no modeling: the **CBOE PutWrite and BuyWrite indices**, a fund's actual dividend-adjusted total return, real **OptionsDX end-of-day option chains** (bid/ask fills, real deltas), and **Yahoo** daily prices and dividends. Where we must model a premium, we price it with the same Black-Scholes engine behind this site's calculators using VIX as the volatility input, and we state which way that biases the result. Every study names its window and its assumptions, the correctness-critical figures are re-derived by an independent script. The headline results on this page are pulled live from those data files; the labels and explanation are editorial copy checked against each study. They are educational, not advice. The bottom line Across 9 backtests on real market data, no options-income strategy beat simply holding the index on total return - every one traded return for a smoother ride, and high win rates did not guarantee strong returns. The durable edge from selling options is lower drawdown, not higher return; a distribution yield is not a return; and defined risk is insurance you pay for. Judge any income strategy by its worst month and its total return, never its win rate. ## Frequently asked questions Do options income strategies beat buy-and-hold? No - not on total return, in any of our backtests. Real CBOE index and market-chain data put cash-secured puts at 6.9% a year versus the S&P's 10.8%, the Wheel (2010-2023 real fills) at 6.5% versus 12.7%, and monthly short strangles at 0.8% versus a price-return (dividends excluded) 10.7% - on total return the strangle's gap is wider still. Selling options traded return for a smoother ride, not a higher one. Is selling options profitable? Modestly, and less than simply owning the index. Across our studies the premium-selling results ran from near-zero (0.8%) to high-single-digit (8.3%) annual returns, all below their benchmarks. Their durable advantage was risk, not reward: the CBOE PutWrite index cut the S&P's worst drawdown from 55% to 37%. Judge these strategies by drawdown, not by the headline income. Does a high win rate mean an options strategy is safe? No, and this is the most common trap. The far-out-of-the-money 10-delta Cash-Secured Put won 95% of months yet made just 1.3% a year. It returned -19.8% across calendar 2020, while its worst month was -25.1%; those are different measures. A 5-delta Short Strangle won 96% of months yet made only 0.7% a year, and one crash erased 110 average winners. Win rate measures how often you win, never how much. What is the real benefit of selling options if it underperforms? Lower drawdown and a smoother equity curve. Every income strategy we tested gave up total return but fell less in crashes: put-writing dropped a maximum 37% against buy-and-hold's 55%, and on real fills the Bull Put Spread's own worst month (-6%) ran far shallower than the naked Put's (-29%) over the same window. If you value sleeping through a crash more than maximizing return, that trade can be worth making. Are covered-call ETFs like QYLD worth it? Only if you understand a distribution is not a return. QYLD paid a headline ~14% yield yet its share price fell 28% since 2013, and its total return (+172%) badly trailed the Nasdaq-100 it writes Calls on (+840%). The high monthly payout is partly your own capital handed back - fine for cash flow, poor for growth. Which options income strategy is best? There is no single best - and chasing an "optimal" setting is usually fitting noise. Our Short Strangle sweep tested 7 strike deltas and found them statistically indistinguishable (the whole cross-strike spread was smaller than the error on any one). Pick by the risk you can hold: defined-risk spreads cap the catastrophe, and put-writing cut the index's worst fall from 55% to 37%. All still lag buy-and-hold on return. Is the Wheel strategy profitable? Yes, but less than buy-and-hold. Over 2010-2023 (14 years) of real OptionsDX fills on SPY the full Wheel - sell a Put, take assignment, sell Calls, repeat - earned 6.5% a year against buy-and-hold's 12.7%, at a 29% maximum drawdown, holding shares about 25% of the time - 51% of holding's annualised return (on cumulative return the gap is wider still). The study records no buy-and-hold drawdown on the same real-fill basis, so we do not claim the Wheel was the smoother ride over this window. How reliable are these backtests? Every figure is computed from real market history - published CBOE index series, real fund total returns, and real OptionsDX end-of-day option chains - and regenerates from the underlying data, with the correctness-critical results independently re-derived. The main caveat is common to all backtests: each is one historical window on mostly one underlying, held mechanically, so treat the direction as robust and the exact number as sample-specific. Educational, not advice.