Metrics & Parameters Glossary

The annualized growth rate of your investment over the backtest period. It represents the constant yearly rate that would yield the same total return.

Formula: CAGR = (Ending Value / Starting Value)^(1 / Years) - 1

A CAGR of 15% means your portfolio would grow by 15% per year on average. This is the most important metric for comparing long-term strategy performance.

The cumulative percentage return over the entire backtest period.

Formula: Total Return = (Ending Equity / Starting Equity) - 1

A total return of 150% means your initial investment increased by 150%, so 100,000 became 250,000.

The absolute dollar amount gained or lost over the backtest period. This is simply the difference between ending equity and starting equity.

While useful for understanding absolute gains, PnL doesn't account for risk or the time value of money, so CAGR and Sharpe are better for strategy comparison.

The largest peak-to-trough decline in equity during the backtest period. This measures the worst loss you would have experienced from a previous high point.

A max drawdown of -25% means at one point, your portfolio declined 25% from its highest value. Lower drawdowns indicate more stable strategies.

Drawdown tests your risk tolerance. A strategy with 50% annual returns but -60% drawdown might be psychologically unbearable.

A risk-adjusted return metric that measures excess return per unit of risk (volatility).

Formula: Sharpe = (Mean Return - Risk-Free Rate) / Std Dev of Returns

Sharpe < 1.0: Poor | 1.0-2.0: Good | 2.0-3.0: Very good | > 3.0: Exceptional

The total number of round-trip trades (buy and sell pairs) executed during the backtest period.

More trades mean higher transaction costs. A strategy with 500 trades vs. 50 trades needs to generate enough extra return to justify the additional costs.

The percentage of trades that were profitable.

Formula: Win Rate = (Winning Trades / Total Trades) x 100

Win rate alone is misleading. A strategy with 90% win rate but small wins and large losses can still lose money. Always consider win rate alongside average win/loss size.

+1 (BUY): Strategy indicates a long position should be taken.

0 (HOLD): Strategy indicates holding current position or staying in cash.

-1 (SELL): Strategy indicates closing positions or taking a short position.

The final stock signal is computed by combining individual strategy signals using either majority vote or weighted vote (typically CAGR-weighted).

The minimum number of time periods (bars) that must pass between consecutive trades. Prevents overtrading and reduces transaction costs.

With cooldown_bars=5 on daily data, after selling a position, the strategy must wait at least 5 days before buying again.

A momentum oscillator ranging 0-100. RSI > 70: overbought (sell). RSI < 30: oversold (buy).

Parameters: rsi_window (14), rsi_overbought (70), rsi_oversold (30).

Trend-following momentum indicator. Buy when MACD line crosses above signal line; sell when it crosses below.

Parameters: macd_fast_window (12), macd_slow_window (26), macd_signal_window (9).

Volatility bands above and below a moving average. Price at lower band = oversold. Price at upper band = overbought.

Parameters: bb_window (20), bb_k (2.0 standard deviations).

Average price over N periods. Price > SMA = bullish. Price < SMA = bearish.

Parameters: window (50 or 200 commonly).

Multi-line trend system with support/resistance cloud. Signals based on Tenkan/Kijun crosses and cloud position.

Parameters: tenkan (9), kijun (26), senkou_b (52), shift (26).

Percentage of available cash to use when buying (1.0 = 100%, 0.5 = 50%).

Percentage of held shares to sell when selling (1.0 = 100%, 0.25 = 25%).

Minimum return threshold before allowing a sell (prevents premature exits).