Every Monday, if flat, buy 100 AAPL shares at the print and write one Friday-expiring call. The strike is chosen by solving for the listed strike closest to a target delta — the target itself shifts with the current implied-vol regime (further OTM when vol is rich, closer to the money when it's calm) and pulls in further after a losing week, so a stock drop gets more room to recover before the position risks getting capped again. Friday: OTM expires (keep the shares, keep the premium, write the next week's call without re-buying); ITM is assigned (shares sold at the strike, flat, restart Monday).
Midpoint assumption
Every fill in this backtest is taken at mid = (BID + ASK) / 2. To justify that, every hourly bar with a real trade print across the strikes actually used is plotted here: TRDPRC_1 (last trade) against mid at that same bar.
Blotter
Every trade actually booked — no working orders, no signal chart standing in for a blotter. Friday OTM is EXPIRE; Friday ITM is ASSIGN on the call plus a STOCK SELL at the strike.
| Time | Instrument | Side | Qty | Limit | Fill | Cash Δ | Notes |
|---|
Ledger
Running position, marked at each week's close. Cash only moves on a blotter event: buying stock, collecting premium, expiring at $0, or assignment at the strike. Short Call shows the strike written that week and its expiry, once it's resolved. Excess equity (NAV minus maintenance margin) is the margin-call line — if it goes negative, the account would be in a margin call.
| Week of | Shares | Short Call | Spot | Cash | LMV | NAV | Initial | Maint. | Available | Excess |
|---|
Write-up
How the strike is chosen
To choose the strike, I wanted to find a rule that was actually used in the real world, so I went with the delta targeting strategy. The way this works is that instead of picking a random constant strike, I am targeting a Black-Scholes delta; this delta is chosen based on market volatility and last week's result. If the market is very volatile, then we will choose a lower delta because the premium is already richer even further out; on the other hand, if the market is quite calm, then the delta chosen is higher, since we need to sit closer to the money to collect a meaningful premium at all. As well, the delta is also affected by last week's performance: after a week when the stock lost money, we decrease the delta. Selecting a lower delta target sets the strike price deeper out of the money, providing additional buffer before the portfolio's gains face capping.
Wait through expiry
In this approach, there are no rolls and no buy-to-close trades. The short call option simply runs through expiration: it either finishes out of the money (allowing us to keep the full premium alongside the underlying stock before writing a new contract) or gets exercised (where our shares are called away at the strike price, leaving us neutral until resuming the strategy next week).
Fill assumption
For execution, each order is filled at the midpoint price, calculated as (BID + ASK) / 2 recorded at that specific moment. In the chart above, this approach is evaluated by comparing it directly to actual transaction data for the strikes selected during backtesting.
Why Reg T, not portfolio margin
Reg T is an easier-to-compute formula and more reliable without having access to a broker's specific risk management book. With Reg T, we are sure to be in the legal limit without making any guesses.
Could this trade actually have been put on?
With $10,000 of starting capital, most weeks in this backtest show negative available funds (see the Ledger above) — buying ~$30,000 of AAPL stock requires roughly $15,000 of initial margin at Reg T's 50% rate, more than double what this account started with. In practice, that means: as capitalized here, this account could not actually have put most of these trades on. A real version of this strategy would need starting capital closer to 1.5× the worst-case initial margin in the window (roughly $15,000–$17,000 for AAPL over this period) to clear margin comfortably every week rather than most weeks.
Known limitation
This backtest only checks the value of the position twice a week: once on Monday right after buying stock and writing the call, and once on Friday after the call is resolved. In between those two days, the option's real market price is changing constantly: it moves with the stock, with time passing, with volatility, but we don't track any of that. So the NAV chart is accurate at the start and end of each week, but it can't show what happened in the middle; if the stock spiked on a Wednesday and came back down by Friday, that swing simply wouldn't show up. With this exact algorithm, it does not hurt much because we would not be doing any actions during the week anyways, but if we were to be able to change our position during the week, then we would have to change this.
The strategy's strike choice also depends on comparing today's volatility to its own recent history. But at the very start of the backtest, there is no "usual" yet to compare against. The first few weeks are judging volatility off just one or two data points, which isn't a very reliable comparison. By the later weeks, once more history has built up, that comparison becomes much more meaningful. So the earliest strike choices in the backtest are a bit shakier than the later ones, simply because the model hadn't warmed up yet.
When does the strategy work, or hurt?
Running the identical strategy across 21 stocks over the same 13-week window (see the separate Regime Analysis page) shows a real, moderate negative correlation (r = -0.51) between a stock's own buy-and-hold return and how much the covered call strategy gives up relative to it. In stocks that fell or stayed flat, the strategy consistently matched or beat simple stock ownership: the collected premium cushioned the decline, and getting assigned was rare, so most weeks kept full exposure to any recovery. In stocks with the strongest run-ups, the strategy gave back the most ground, since more weeks ended in assignment and capped the gain right at the strike.
Two results don't fit that pattern cleanly, and are worth naming rather than hiding. XOM rallied +18.5%, yet the strategy still slightly outperformed (+23.3%). IBM is another outlier: the stock was down only -4.7%, yet the strategy lost -14.9%, a -10.2-point gap in the wrong direction. IBM was assigned 7 of 13 weeks, which is far more than any other name, meaning it wasn't a smooth decline but a choppy one, repeatedly rallying just enough to trigger assignment before falling again. That pattern (frequent short-lived rallies inside a longer downtrend) is close to the worst case for this strategy: it caps the rallies on the way up without ever getting the benefit of riding a sustained recovery.