Methodology

1. What We Track

We identify and track investment signals from newsletter author content using our proprietary analysis methodology. A signal is identified when our analysis interprets directional conviction about a specific ticker from an author's published content.

We focus on thesis-level signals — not passing mentions or comparative analysis. When an author expresses conviction that a stock will outperform or underperform, that's a trackable signal.

2. Pick Definition

Picks are categorized into three levels:

• Thesis: Explicit directional call with supporting analysis. Opens or reinforces a position.
• Mention: Substantive discussion with price targets or key metrics, but not a clear position call.
• Passing: Brief reference without investment thesis. Not tracked for performance.

Only thesis-level signals create or modify positions. Mentions and passing references are captured for context but don't affect performance metrics.

3. Return Calculation

Returns are calculated using dividend-adjusted close prices:

• Entry price: Adjusted close on the publication date of the initial thesis signal
• Exit price: Adjusted close on the date of sentiment reversal, or the last available price for open positions
• Return %: (Exit - Entry) / Entry × 100 for bullish positions; inverted for bearish

We use adjusted close prices from Tiingo to properly account for splits and dividends.

4. Benchmark Selection

Each author is assigned a default benchmark based on their coverage focus:

• Tech/Semiconductor: SMH (VanEck Semiconductor ETF)
• Energy/Commodities: XLE (Energy Select Sector SPDR)
• Broad Market: SPY (S&P 500 ETF)

Alpha is calculated against this benchmark over the same holding period.

5. Alpha Definition

Alpha measures outperformance versus the benchmark:

Alpha % = Position Return % − Benchmark Return %

A positive alpha means the pick outperformed the relevant sector; a negative alpha means it underperformed. This helps distinguish skill from simply riding a bull market.

6. Win Rate

Win rate is the percentage of closed positions with a positive return:

Win Rate = (Positions with Return > 0) / (Total Closed Positions) × 100

Note: A high win rate with small wins and large losses can still result in negative overall returns. Always consider win rate alongside average return.

7. Sample Size & Confidence

Statistical confidence increases with sample size. We use a 3-tier system:

• < 10 closed positions: "Insufficient data" — not ranked on leaderboard
• 10-29 positions: "Early data" — ranked with caveat that results may be noisy
• 30+ positions: "Established track record" — high confidence in statistical validity

We also display confidence intervals (±CI) on average returns to show the range of likely true performance.

8. Cross-Author Convergence

When multiple newsletter authors independently reach the same conclusion about a stock, it may indicate a stronger signal. We track these cross-author agreements and disagreements.

Convergence (Agreement)

Convergence occurs when 2+ authors have the same directional view (bullish or bearish) on a ticker within a rolling window. We calculate:

• Strength: Based on the number of agreeing authors. 2 authors = Moderate, 3+ = Strong.
• Time window: 60-day rolling window by default. Authors must have active positions within this period.

Divergence (Disagreement)

Divergence occurs when authors have opposing views on the same ticker — one bullish, one bearish. These "contested picks" highlight areas of genuine disagreement among experts.

• We display which authors are bullish vs bearish
• Recent thesis changes (upgrades, downgrades) are highlighted
• Divergence does not indicate which view is correct — it shows active debate

Note: Convergence is informational only. Multiple authors agreeing does not guarantee the thesis is correct. Always conduct your own research.

9. Position Rules

Positions are subject to the following rules:

• Minimum conviction: Multiple thesis-level signals are required to open a position
• Minimum holding period: Positions must be held for a minimum period before a reversal signal can close them
• Auto-close: Positions with no new signals for an extended period are automatically closed at the last available price
• Author-specific parameters: Tracking parameters (signal types, holding rules) are calibrated per author based on their publishing style

10. Limitations

• Signal identification is performed through our proprietary analysis methodology. This process involves interpretation and may differ from the author's actual recommendations or portfolio holdings.
• Entry prices use the publication date, which may differ from when the author formed their view.
• With fewer than 30 closed positions per author, statistical measures have wide confidence intervals.
• Performance data is hypothetical and does not represent actual trading. It does not account for transaction costs, slippage, or taxes.
• Alpha is calculated as the arithmetic difference between position return and benchmark return.

11. Disclaimers

• We do NOT provide investment advice or recommendations
• We do NOT guarantee that past performance predicts future results
• We do NOT endorse any newsletter author or their strategies
• We do NOT track options, futures, or other derivatives

Author names are used for identification purposes only. SavantAlpha is not affiliated with, endorsed by, or sponsored by any tracked author or their publications. Performance metrics represent our independent interpretation of published content and may not reflect the author's actual positions or results.

SavantAlpha is a performance tracking service. All investment decisions should be made with the guidance of a qualified financial advisor. See our full disclaimer for more information.