Risk is what remains after confidence is removed.

Volatility-managed equity exposure, tested against three decades of data.

2024—26INDEPENDENT RESEARCHERPYTHON · PANDAS · BACKTESTINGSHARPE 0.94 VS 0.61 ※

THE SYSTEM UNDER TEST — SHOCK / DAMPED · PROCEDURAL

THE 45-SECOND READ

PROBLEM

A persuasive backtest curve is not evidence until the route to it can be inspected.

DECISION

Scale exposure by realized volatility instead of forecasting returns.

METHOD

A volatility-targeted overlay tested against three decades of equity data, with costs and failure regimes in the foreground.

STATUS

Sharpe 0.94 vs 0.61 baseline — draft figures; methodology pending review.

THE TURN

Prediction became exposure.

The draft turns away from the clean forecast and toward behaviour under pressure: what the method does when conditions stop flattering it.

THE PRACTICE / 01

Stress first. Admire later.

Assumptions, costs, data lineage, and failure regimes belong in the foreground. A persuasive curve is not evidence until the route to it can be inspected.

Equity curve — notebook figure

EVIDENCE / 02 — DROP REAL CAPTURES HERE

GENERATED IMAGERY IS NEVER USED AS EVIDENCE

Drawdown / regime comparison

FIG. 02 — BEHAVIOUR UNDER STRESS

Data provenance table

FIG. 03 — DATA PROVENANCE

EVIDENCE LEDGER / 03

What the working draft shows.

— A volatility-targeting rule with explicit cost and turnover accounting— Backtests across three decades of daily equity data— Factor-model context for where the effect does and does not hold

What this page does not claim yet.

— Sharpe figures are draft placeholders, pending an audited backtest— Notebooks, code, and data provenance are not yet published— Real figures still need to be dropped into the evidence slots above