The performance

Against the market

The same account, three ways. Portfolio Full, SPY and QQQ each start with $451,350 on 2 May 2022 — the most margin the Portfolio ever needed — and run to 30 October 2026. While that money sits as margin it earns the T-bill rate; the index funds reinvest their dividends.

Researched in-sample, tested out-of-sample

The rules were developed on the window of 2 May 2022 to 30 April 2026: that is the in-sample period that was used to derive the rules that work. The crucial test to avoid curve-fitting is if these rules keep working in a different period that was not used to develop them. From 1 May 2026 the same rules are tested on data that played no part in researching them — the out-of-sample period, published as the official execution record. It is priced more harshly too: the record’s fills allow two minutes for a subscriber to act after each alert; the backtest’s do not.

Out-of-sample, the Portfolio held up: a Sharpe ratio of 3.83 against 3.63 in-sample, a shallower deepest drawdown (−⁠8.9% of the account against −⁠12.0%) and a smaller worst day (−⁠6.4% against −⁠10.4%). One row moved the wrong way: correlation to SPY was 0.33 out-of-sample against 0.08 in-sample. Still low, but higher. And 127 sessions is a short record: one bad day can change these numbers a great deal, which is why the out-of-sample return is shown for the period and not annualised.

Results from 1 May 2026 onward are simulated, not broker-executed. Fills are modelled from 1-minute quote data two minutes after the time stamped on the alert — an allowance for the time it takes a subscriber to act — with strikes re-selected from the option chain at that moment. Commission of $1.20 per leg is deducted.

The whole period (May 2022 – Oct 2026)

All three sizes against the indices

Going down the sizes, the annual return rises (41.2%, 52.7%, 59.2%) while the Sharpe ratio falls (3.64, 3.07, 2.28): the smaller sizes earn more on their capital by taking more risk with it.

What this comparison flatters, and what it doesn’t

• The account is the peak margin. Each Portfolio size is measured against the most margin it ever needed across the whole period ($451,350 at Full; the out-of-sample peak sat slightly above the backtest’s, so the account here is a little larger than the one on the Portfolio page). On an average day far less was in use but you need to be sized for the peak days.
• It is assumed that cash is used for margin that sat idle and earned the 3-month T-bill rate (about 4.4% on average over the window). That interest is included here; the Portfolio page, which covers the backtest alone, shows returns without it.
• The index is fully invested; the Portfolio never grows its size. SPY and QQQ compound every dollar. The Portfolio trades one contract per leg from start to finish, however large the account becomes. That is conservative for the Portfolio.
• The window matters. It starts in the 2022 bear market, which deepens the index drawdowns, and then covers a strong bull market, which lifts the index returns. Both are shown as they happened.
• Two periods, one account. The charts and the table above run one account through both periods so you can see the whole experience; the shading marks where the backtest ends. The in-sample and out-of-sample table measures each period on its own, so neither borrows from the other.
• All trades are simulated net of the full bid-ask spread and $1.20 commission per leg. The index figures are actual. Sharpe and Sortino ratios are measured on returns above the T-bill rate for all five columns.

Inside the day

A day’s closing result hides what the day felt like. Here every open position was marked every minute at what it would have cost to close it — buying back the short leg at the ask and selling the hedge at the bid — for all 1,071 sessions: 944 in the backtest and 127 since. This is what you sit through between the open and the close.

Did the days feel different out-of-sample?

Not much. The out-of-sample days went below zero just as often and stayed there about as long; the typical hole was a little deeper and the worst one shallower. The record is still short, so its deepest point says less than the backtest’s.

Two deep holes, opposite endings

The two deepest intraday holes of the whole period that ended in opposite ways. The depth of a hole tells you very little about how the day will close, which is exactly why closing positions outside the rules is the expensive mistake.

Running P&L Portfolio Full, marked every minute from 9:30 to 16:00 ET. Both days are in-sample: hypothetical / backtested.

How deep the hole went, and how the day ended

A shallow dip is routine and usually recovers. The deeper the hole, the less often the day comes back, but even past $10,000 more than one day in five still closed green.

Why four strategies are calmer than one

Run separately over the whole period, the four strategies’ worst drawdowns add up to −⁠$154,263 at Full. Run together, the Portfolio’s worst was −⁠$55,572: 36% of that sum, and shallower than the worst of any one strategy on its own. When one strategy is in its worst stretch, the others usually are not, which is the case for running all four rather than picking one.

All three sizes - Intraday

Smaller sizes go less far below zero in dollars, but further as a share of the account they need: the same pattern as their drawdowns.