This post collects backtest statistics and investment metrics I have learned over the years which I used to evaluate investment strategies. A link to this Jupyter Notebook shows the implementations with Python with NumPy and Pandas.

General Characteristics

• Time Range
• Average AUM
• Capacity
• Leverage
• Maximum Dollar Position
• Ratio of Longs
• Frequency of Bets
• Average Holding Period
• Annualized Turnover
• Correlation to Underlying

Performance Metrics

• PnL
• PnL from Long Positions
• Annualized Rate of Return Time-Weighted Rate of Return
• Average Return from Hits
• Average Return from Misses

Run

• Returns Concentration
  • High sharpe ratio
  • High number of bets per year
  • High hit ratio
  • No fat tail
  • Bets are not concentrated in time
• Drawdown and Time under Water
• Runs Statistics for Performance Evaluation
  • HHI index on positive returns
  • HHI index on negative returns
  • HHI index on time between bets
  • 95-percentile DD
  • 95-percentile TuW

Implementation Shortfall

• Broker fees per turnover
• Average slippage per tunrover
• Dollar performance per turnover
• Return on execution costs

Efficiency

• Sharpe Ratio
• Probabilistic Sharpe Ratio
• Deflated Sharpe Ratio
• Information Ratio

Classification Score

• Accuracy
• Precison
• Recall
• F1 Score
• Negative log-loss

Attribution

• Barra’s multi-factor

Risk Metric

• Volatility Risk
  • Standard Deviation
  • Downside Deviation
  • Sharpe Ratio
  • Sortino Ratio
• Benchmark Risk
  • Excess Return
  • Batting Average
  • Up Capture
  • Down Capture
  • Alpha
  • Beta
  • R-squared
  • Tracking Error
  • Treynor Ratio
  • Information Ratio
  • M-squared
• Capital Preservation Risk
  • Maximum Drawdown
  • Pain Ratio (Pain Index)
  • Calmar Ratio
• Tail Risk
  • Skewness
  • Kurtosis
  • Omega
  • VaR
  • CVaR