Our Backtest Results: Methodology and What the Numbers Mean

#What We Tested

Our backtest simulates a simple, implementable strategy:

Why These Parameters?

• Top 20 stocks: Concentrated enough to have meaningful factor exposure; diversified enough to reduce single-stock risk. See our portfolio construction guide.
• Equal-weight: Simpler, more robust, and historically outperforms complex weighting schemes (DeMiguel et al., 2009).
• Monthly rebalance: Captures factor signal decay while keeping turnover manageable. See our rebalancing guide.
• 10 bps transaction costs: Accounts for bid-ask spreads on liquid, large-cap stocks. This is conservative — many investors pay less.

#Time Period: January 2022 to Present

Our backtest covers January 2022 through the present — approximately four years of data.

What Happened During This Period

This period includes a bear market, a narrow rally, a broadening rally, and rate uncertainty — a reasonable test of different market regimes, though not a complete market cycle.

Why Not a Longer Backtest?

We deliberately chose a shorter, more honest period:

We will extend the backtest as more out-of-sample data accumulates. Every additional month strengthens the evidence.

#The Results

Cumulative Performance

Risk-Adjusted Metrics

What These Numbers Mean

Sharpe ratio of 1.05: For every unit of risk taken, the portfolio earned 1.05 units of return. A Sharpe above 1.0 is considered excellent for an equity strategy. For context, most hedge funds target a Sharpe of 0.5-1.0.

Max drawdown of -18.7%: The worst peak-to-trough loss was 18.7%. This occurred during the Q1-Q2 2022 bear market. The S&P 500 drew down 25.4% over the same period — meaning our portfolio lost less in the worst environment.

Alpha of +8.9%: After adjusting for market exposure (beta of 0.88), the portfolio generated 8.9% of excess annual return. This is the return attributable to factor selection, not simply being invested in stocks.

Beta of 0.88: The portfolio is slightly defensive — it captures about 88% of market movements. This is consistent with the low-volatility factor tilt.

#How Returns Are Calculated

Point-in-Time Data

The most critical aspect of any honest backtest: we use point-in-time data. This means:

• On January 1, 2023, the model only sees data that was actually available on January 1, 2023
• No future earnings, no restated financials, no data that would have been unknown at the time
• Financial data uses the most recent quarterly filing as of each rebalance date

This eliminates look-ahead bias — the most common source of inflated backtest returns.

Return Calculation

Monthly returns are calculated as:

1. 1At the start of each month, select the top 20 stocks by composite score
2. 2Equal-weight the portfolio (5% each)
3. 3Let the portfolio run for one month (no intra-month trading)
4. 4At month-end, calculate the portfolio return including dividends
5. 5Subtract estimated transaction costs (10 bps for buys and sells)
6. 6Repeat

Survivorship Bias

We include all stocks that existed in our universe at each point in time, including stocks that were subsequently delisted or acquired. This prevents survivorship bias — the tendency for backtests to look better because failed companies are excluded.

#Limitations and Caveats

We believe in radical transparency about what our backtest does and doesn't prove.

What the Backtest Shows

• Our six-factor composite score identified stocks that outperformed the S&P 500 over the test period
• The outperformance is consistent across different market environments within the test period
• Risk-adjusted returns (Sharpe, Sortino) are strong, not just raw returns

What the Backtest Does NOT Show

Honest Caveats

1. 1The test period is favorable. A strategy that weights profitability and momentum heavily would have struggled more in 2019-2020 when speculative, unprofitable stocks led the market.

1. 1We chose the parameters. Even though our weights and thresholds are based on academic literature (not tuned to this backtest), the choice of which factors to include is itself a degree of freedom.

1. 1Transaction costs are estimated. Real trading costs depend on your broker, order size, market conditions, and execution quality. Our 10 bps estimate is reasonable but not exact.

1. 1Tax effects are excluded. Monthly rebalancing generates short-term capital gains, which are taxed at higher rates. In a taxable account, after-tax returns would be lower.

#How to Use These Results

Do: - Use the backtest as **one input** in your investment decision, not the only input - Focus on the **risk-adjusted metrics** (Sharpe, Sortino, max drawdown), not just the raw return - Understand that the **process** (systematic factor selection) is more important than any specific return number - Expect periods of underperformance — even the best strategies have bad months and bad quarters

Don't: - Assume +19.4% will repeat every year - Invest money you'll need in the next 3-5 years based on backtest results - Ignore the limitations listed above - Compare this backtest to strategies tested over different time periods

#The Bottom Line

Our backtest demonstrates that a simple, transparent, six-factor strategy has outperformed the S&P 500 over the past four years with lower risk. The methodology is sound: point-in-time data, survivorship-bias-free, with realistic transaction costs.

But four years is a starting point, not proof. We will continue to publish updated results as more out-of-sample data accumulates. The ultimate test of any investment strategy is time — and we're committed to transparency every step of the way.

See the latest rankings →

#Further Reading

• How Composite Scores Work →
• Why We Use Exactly Six Factors →
• Sector-Relative Scoring Explained →
• How to Build a Portfolio Using Our Rankings →
• Our Complete Methodology →