Backtested: What 7 Years of Option Data Taught Us About Selling Premium

The machine we built

Most option-selling advice you will find online comes from one of two places: a backtest the author will not show you, or a memory of what worked for them in one specific market. We wanted something sturdier, so we built our own backtest harness — a program that replays simple, rules-based option strategies on seven years of real end-of-day option prices, from January 2019 through 2026.

Seven years is not an accident. It covers four very different regimes:

Every trade the harness runs is recorded, win or lose. When we publish the verdicts, you will see the losers alongside the winners. That is the deal.

What we are testing — not what we are claiming

Note the verb: testing. The full verdicts are still being finalized. What we can tell you honestly today is the method and the questions.

We are testing DTE and delta choices across the four strategies we actually trade: cash-secured puts, put credit spreads, short strangles on futures, and covered calls. For each one, the harness asks the same blunt question: which entry settings actually earned their keep, mechanically, across all four regimes?

That means we hold everything else fixed — the same mechanical rules, the same profit targets, the same exits — and vary one thing at a time: how far out we sell, how far from the money we strike, when we take profits, when we cut losses. A setting that only works in calm bull years is not a setting. It is a lucky draw.

What our own 2026 tape already showed us — labeled preliminary

While the seven-year harness grinds, we also have our own live trading history from 2026. It is real money, real fills, every trade tracked — but it is also one regime year, mostly a bull market with bursts of volatility. So read these as preliminary signals, not conclusions.

Signal 1: 40–50 DTE was the sweet spot on our futures tape

On our 2026 futures short-strangle tape, entries around 40–50 days to expiration were the most consistently profitable — an 88% win rate in that window on our tracked trades. Preliminary, and concentrated: the bulk of that profit came from index products (NQ and ES), so it may be telling us as much about which underlyings we traded as about the DTE choice itself. The seven-year test will show whether 40–50 holds up outside one product mix and one year.

Signal 2: our early exits may have cost us roughly forty to forty-seven thousand dollars — modeled

This one stung. We compared our actual 2026 exits — the human ones, taken early because the screen looked scary — against a mechanical rule: take profit at 50% of premium, exit at 21 DTE regardless. The mechanical rule comes out ahead by a modeled $40,000–$47,000 on our 2026 tape. Labeled plainly: that is a model result on one year of our own history, not a promise, not a claim that a robot beats a trader in all years. But it is the kind of number that makes you write down your exits in advance.

Signal 3: systematic covered-call overwriting dragged in calm years

We backtested a plain systematic covered-call overwrite — selling calls against a stock position month after month, no discretion — and in calm, upward-drifting years it lagged simply holding the stock by roughly 6% to 15% per year. In the 2022 bear year it helped, adding roughly 11% to 15% by cushioning the fall. Our working verdict: keep covered calls as a tactical tool, not a default habit. Also preliminary, also subject to the full seven-year run.

Why we publish the method before the verdicts

Because nobody else in retail options education does. The usual format is: guru states rule, guru sells course, backtest never appears. We are doing it the other way around. Here is the rig, here is the data span, here are the questions, here are the preliminary signals with their labels, and when the full verdicts land, this page will be updated — winners, losers, and the trades in between.

We would rather show you how the sausage gets tested than sell you the sausage recipe on vibes.

How to read what comes next

A few ground rules while the verdicts are pending:

• Preliminary means: seen on our 2026 tape or a partial run, not yet validated across regimes.
• Modeled means: a calculation on historical data, not a guarantee about the future.
• If the seven-year run overturns one of these signals, we will say so on this page. That is the point of the exercise.
• Nothing here changes how we trade day to day. Our daily screens run on rules we trust now; the backtest is the audit, not the engine.

Check back. The verdicts are coming — and the losers are coming with them.

The four regimes, briefly

A backtest is only as honest as the markets it includes, so here is what each regime actually felt like for a premium seller:

• 2019: Volatility sat near historic lows for months. Premiums were thin, win rates were high, and the danger was boredom — the temptation to sell closer to the money to juice returns. The harness will show whether that temptation paid.
• Early 2020: The fastest 30% drawdown in history. Short puts got run over; short calls printed money for about three weeks and then got run over on the snapback. Any strategy that claims to handle volatility has to survive March 2020 on the tape, not in a footnote.
• 2022: The slow bleed. No single crash to point at, just month after month of down-drift with volatility spikes that punished anyone selling too close. This is the regime most retail backtests quietly exclude, because it makes everything look bad. We kept it in on purpose.
• 2023–2026: The bull market with tantrums — strong trends upward, interrupted by volatility bursts that tested every exit rule we had. Our own 2026 trading lived entirely inside this regime, which is exactly why we refuse to generalize from it alone.

What we do with the losers

When the verdicts land, the losing trades come with them. Not summarized, not smoothed — shown. A backtest that hides its losers is an advertisement, and we are not running advertisements.

Here is a preview of the honesty standard, from our own 2026 tape: not every 40–50 DTE entry won, our early exits demonstrably cost us modeled money, and there were stretches where doing nothing would have beaten doing something. We publish those stretches because the lesson is in them. Anybody can show you a winning trade. We are trying to show you which rules survive contact with all four regimes — and the rules that do not survive are arguably the more valuable finding.

That is the whole bet behind this project: that showing our work, losers included, across seven years and four regimes, is worth more than any single verdict. The verdicts are coming. The method is already here.