“Sell in May and Go Away”: The Honest Test We're About to Run
"Sell in May and Go Away": The Honest Test We're About to Run
Every month we take one popular Wall Street rule and test it mechanically — no cherry-picking, no curve-fitting. Next up: "Sell in May and go away." The claim: stocks do all their work in the winter half of the year, so you're better off holding November through April and sitting in cash May through October.
Before we run it, we're publishing the test design. That's deliberate — deciding the rules after seeing the answer is how backtests get gamed.
The mechanical test
- Strategy: Hold SPY from the close of the last trading day of October through the close of the last trading day of April. Sit in cash May through October. Repeat every year.
- Benchmark: Buy-and-hold SPY over the identical window.
- Ticker: SPY. Data: daily adjusted closes (dividends included via the adjustment).
- Execution rules: our standard published rules — signal at the close, earning starts the next bar, no look-ahead, exit-day returns counted.
What would count as "confirmed"?
The seasonal strategy needs to beat buy-and-hold on total return, not just on vibes. We'll also report max drawdown and time in market, because a strategy that returns slightly less with dramatically smaller drawdowns is a different — and legitimate — kind of win.
What could make this test unfair
- Transaction costs: The strategy trades twice a year — trivial. Costs won't decide this one.
- The cash assumption: Sitting in "cash" historically earned T-bill interest. We'll disclose whether we credit it (we lean toward crediting short-term Treasury rates — cash isn't a mattress).
- Start-date sensitivity: Seasonal effects are notoriously period-dependent. We'll note how sensitive the result is to the chosen window.
The honest priors
"Sell in May" has genuine historical support in some markets and periods — the winter-half outperformance of equities is one of the better-documented seasonal anomalies. But anomalies decay once published, and the last decade of mega-cap dominance may have rewritten the pattern. We genuinely don't know which way this one lands.
That's the point. We'll run it, publish the numbers — win, lose, or "it's complicated" — with the same execution discipline as everything else.
Not financial advice. This is a research plan, not a recommendation.