Methodology

The Babeshkin Risk Monitor tracks five independent dimensions of market risk. Each dimension watches for a different type of danger — from rising protection demand to fading momentum to stress in corners of the market that most investors never look at.

When one or two dimensions flag risk, it's usually noise. When four or five flag simultaneously — history says pay attention.

The model doesn't predict where the market is going. It identifies environments where the risk of a significant decline is elevated.

Five Risk Dimensions

Volatility

Monitors the cost of portfolio protection in the options market. When investors expect turbulence, they buy options to hedge — driving protection prices higher. Our model tracks these shifts across multiple horizons and detects when rising demand for protection signals a structural change in market conditions — not just a temporary reaction to headlines.

Market Breadth

Tracks the health of the rally beneath the surface. A rising index can mask underlying weakness — where a handful of large stocks carry the market while hundreds quietly decline. Our model monitors participation across the full index and flags when the rally narrows to a level that has historically preceded corrections.

Credit Stress

Detects early signs of risk aversion in fixed income markets. Credit markets often crack before stocks do. When institutional investors start pulling back from risk, it shows up here first — sometimes weeks before equity markets react.

Convergence

Monitors the structural relationship between short-term and long-term market health. In a strong market, recent momentum builds on a solid foundation. When that relationship breaks down — recent strength fading while longer-term conditions weaken — it often signals that the market is quietly transitioning from expansion to contraction.

Turbulence

Measures the level of disorder across individual stocks in the index. The headline index can appear calm while the stocks inside it churn in unusual ways. Our model analyzes this constituent-level behavior and flags when broad, synchronized stress emerges — the kind that often precedes larger market dislocations.

How the Score Works

Each dimension is either ON (risk detected) or OFF (normal). The score is the count: how many of the five are flagging risk right now.

0-2Low risk. The model stays fully invested.

3Elevated. Something is developing, but not enough to act. Current position is maintained.

4-5High risk. Multiple independent warning signs firing at once. The model exits to protect capital.

Signals are evaluated weekly and executed the following trading day.

What This Model Is Not

This is not a crystal ball. It doesn't predict crashes or pick bottoms.

It's a risk filter — designed to keep you invested during good times and move you out when conditions become dangerous.

The tradeoff is honest: sometimes it exits too early or re-enters too late. Over 20+ years of testing, that cost has been far outweighed by the drawdowns it avoided.

Behind the Numbers

All results are from a rules-based backtest starting in 2003, validated against an independent third-party simulation platform with 100% signal match.

The model was tested using walk-forward validation — training on one period, testing on a completely separate period — to confirm it works in markets it has never seen.

No curve-fitting. No cherry-picking. No lookahead bias.