Performance
Researched results (2015–2025) and live results (November 17, 2025 onward), benchmarked against the S&P 500 (VOO). All data is publicly available for independent verification.
Summary
Values are rounded to one decimal place and reflect the locked V2 baseline computed from the canonical backtest. Past results do not guarantee future performance.
Live results
Live results are tracked separately from November 17, 2025 and updated each trading day. The live dataset uses immutable date-stamped snapshots — each day's result is locked and never overwritten.
As the live track record grows, this section will expand to include live CAGR, drawdown, and edge vs VOO.
What these numbers represent
Performance is computed from the daily signal selections published by the system. Each day's return reflects the theoretical result of holding the selected ETF (and any defensive allocation) for that session, compared to holding VOO.
The researched period (2015–2025) uses historical market data applied to the same rules-based system that runs in production today. Live results (November 17, 2025 onward) are computed from signals published before market open and verified against actual market prices.
This is signal quality measurement — not a portfolio replication track record. The system publishes one lead ETF decision per day. It does not account for taxes, transaction costs, slippage, or individual circumstances.
Verity vs S&P 500
| Metric | Verity | S&P 500 (VOO) |
|---|---|---|
| CAGR (researched) | 22.9% | 12.8% |
| Maximum drawdown | –15.5% | –34.0% |
| Defensive mechanism | Gold + cash overlay in Bear regimes | None (fully invested) |
| Researched period | Jan 2015 – Nov 2025 | |
| Live since | November 17, 2025 | |
How to verify
All performance data is published as machine-readable JSON on the Verity CDN. You can download and audit every data point independently.
-
Performance manifest
manifest.json— dataset inventory with SHA-256 checksums -
Backtest equity curve
backtest_2015_2026.json— daily equity series, 2015–present -
Live performance
live_latest.json— pointer to latest live dataset -
Signal archive
signals_archive.json— every daily selection since 2015 -
Today's signal
latest.json— current day's published signal -
AI declaration
ai.json— machine-readable site metadata
Assumptions
| Signal generation | Uses prior-day closing prices (T) |
| Execution timing | Assumed at next session open (T+1). The system is designed to be directionally robust to normal execution variance. |
| Dividends | Included via adjusted close prices in the source data |
| Transaction costs | Excluded |
| Slippage | Excluded |
| Taxes | Excluded. Note: gold ETFs (GLD) may be taxed as collectibles at up to 28% for long-term holdings. |
| Universe | Fixed set of liquid U.S.-listed ETFs across growth, defensive equity, international, and gold. Reviewed periodically. Context tickers (bonds, dollar, oil) are used as inputs only and never selected. |
| Allocation | K=1 (one lead ETF). In Bear regimes, a defensive overlay splits allocation across equity, gold, and cash based on regime severity. |
| Data immutability | All published datasets are versioned and never overwritten. Each update produces a new file referenced in the manifest with a unique SHA-256 hash. |
Important limitations
Researched results are based on historical simulation. They reflect what the system would have selected using the same rules running today, applied to past market data. They are not actual trading results.
Live results are computed from signals published before market open, using actual closing prices. They reflect theoretical signal quality — not a managed portfolio.
Past performance does not guarantee future results. See the assumptions table above for a complete list of what is and is not included in the calculations.