Summary Statistics
Equity Curve
Starting value $1.00 — weekly rebalancing, equal weight top-5 picks
Simulated returns. No transaction costs, slippage, or taxes included.
How the Algorithm Works
When the S&P 500 is above its 200-day moving average, the algorithm selects the top 5 stocks ranked by a machine-learning model that predicts weekly excess returns. Inputs include earnings momentum, revenue growth, technical signals, and sector context.
When the S&P 500 is below its 200-day moving average, the algorithm switches to a defensive selection model that ranks stocks on stability metrics: low beta, strong balance sheets, and consistent cash flows.
Each week, exactly 5 stocks are selected and weighted equally (20% each). The portfolio is fully rebalanced every week — no compounding of individual positions, no leverage, and no short selling.
Picks are drawn from large- and mid-cap U.S. equities with sufficient liquidity. Financials, REITs, and very small-cap stocks are excluded. The universe resets each week based on current fundamental data.
Weekly Backtested Picks
Top-5 stocks selected by the algorithm each week — most recent first.
| Week | Picks | Return |
|---|
Important Backtest Limitations
- Backtest returns are simulated and do not reflect actual trading. No transaction costs, commissions, or slippage are included.
- Data used to train the model overlaps the backtest period — this introduces look-ahead bias that inflates historical performance.
- Taxes are not modeled. Weekly rebalancing would generate substantial short-term capital gains in a taxable account.
- Results shown assume equal-weight 5-stock portfolios, fully rebalanced each Friday at Friday's closing price.
- The algorithm was designed and tested on historical data — future market regimes may differ significantly.