Current Signals

ML scores and regime analysis as of April 10, 2026.

Current Regime

Crisis

100% confidence

The HMM regime detector classifies the current market environment based on proprietary macro features. Crisis regime triggers defensive positioning across the portfolio. The system is currently rotating out of risk assets and into capital-preservation instruments.

ML Scores

Highest scoring assets — the system is favoring these.

Rank Ticker Category Score
1 GLD Commodity 0.82
2 FXY Currency 0.79
3 SHY Rates 0.77
4 BIL Cash 0.75
5 TLT Rates 0.74
6 IEF Rates 0.71
7 AGG Rates 0.69
8 TIP Rates 0.66
9 SLV Commodity 0.63
10 LQD Credit 0.60

Risk Assets

Lowest Scoring Assets

The system is avoiding these.

Rank Ticker Category Score
1 SPY Equity 0.38
2 QQQ Equity 0.31
3 XLK Equity 0.28
4 IWF Equity 0.26
5 USO Commodity 0.24

Signal Architecture

Signal Categories

Momentum +

Price trend signals evaluated across multiple timeframes. The system measures the persistence and consistency of directional moves for each asset in the universe. Specific signal construction is proprietary.

Volatility +

Risk measurement signals used for position sizing and regime change detection. Elevated or rapidly shifting volatility conditions inform both scoring and the defense doctrine. Specific signal construction is proprietary.

Macro +

Yield curve shape, credit conditions, and monetary policy indicators feed into the macro signal category. These features capture the broader economic environment in which assets are priced. Specific signal construction is proprietary.

Cross-Asset +

Relative strength and sector-level capital flow signals assess how each asset is positioned within the broader market structure. The system identifies capital rotation patterns across asset classes. Specific signal construction is proprietary.

Alternative +

Proprietary signals derived from non-standard data sources and feature engineering processes. These complement the four primary categories and are evaluated on the same factor hygiene standards. Specific signal construction is proprietary.

Methodology

OVRWCH scores every ETF daily using a walk-forward ML ensemble. The model retrains monthly on an expanding window of historical data. It has never scored data it was trained on. Scores represent the model's ranking of expected risk-adjusted performance conditional on the current regime. They are not price targets or return forecasts.

Signal scores are published for informational and educational purposes only. Not investment advice. Scores are model outputs — not recommendations. Past model performance does not guarantee future accuracy. Paper trading since April 10, 2026. Full terms.