← All systems

Harrier_SPY

Long/flat time-series trend overlay

Holds a broad equity index or holds cash, and nothing else. Its purpose is the reduction of drawdown and volatility rather than the addition of return: across the test it delivered roughly half the benchmark’s drawdown for roughly 1.4 percentage points a year less return. Operationally undemanding (fewer than one position change a year) but psychologically severe, because it spends long stretches visibly losing to the obvious alternative in exchange for protection that arrives rarely and cannot be scheduled.

Direction

Long only: fully invested or fully in cash

Universe

One broad US large-cap index proxy, fixed

Holding period

Median 303 days; range 30 to 1,492

Positions held

One at a time, or none

Action needed

Monthly: one check at the month-end close, any order at the next open; a change comes less than once a year

Tooling

Python, end-of-day data

Simulated results only; no live account results appear anywhere in this record. Full available history of a single continuously listed exchange-traded proxy for a broad US large-capitalisation index, total-return adjusted. The instrument is fixed for the whole test, so there is no selection and no survivorship bias in the usual sense, though the underlying index reconstitutes on its provider’s schedule and this study does not isolate that effect. Per-share commission with a per-order minimum; proportional slippage on every fill. State is assessed at a session close and any position change executes at the next session’s open, so no result can act on same-bar information. Compounding single account, no additions or withdrawals; idle cash credited at a published short-term government rate, which a real account would not fully receive. CAGR and volatility from the daily series; maximum drawdown measured on daily marks (−24.4%); recomputing it from the monthly curve below gives a shallower −21.8%, because monthly sampling cannot see intra-month troughs. Sharpe is on an excess-return basis net of the credited rate; the zero-rate figure would be 0.77. No correction for multiple comparisons: roughly 300 configurations were examined, though this one was specified in advance rather than selected from them. Benchmark on identical assumptions: CAGR 10.84%, max drawdown −55.2%, volatility 18.5%.

← All systems