Statistical Testing and Validation of Investment Ideas Flashcards
Key Statistical Measures in Finance
- Mean
- Variance and sd dev
- Skewness
- Kurtosis
- Correlation and Covariance
- Beta
- Sharp ratio
- Value at Risk
Importance of Statistical Measures in Finance
- Risk Assessment
- Investment Evaluation
- Portfolio Construction
- Performance Benchmarking
Challenges and Considerations
- Non - stationary
- Model assumptions
- External Factors
Essence of Hypothesis testing
- Null Hypothesis (H0)
- Alternative Hypothesis
- Test Statistic
- Significance level (alpha)
- P-value
Applications in Finance
- Teste asset returns
- Evaluating Investment strategies
- Cointegration teste
- Event studies
- Testing Market Efficiency
Steps in Hypothesis testing
- Formulate the Hypotheses
- Choose the significance level
- Select the Appropriate Test
- Compute the test statistic
- Make a decision
Challenges and Considerations in Financial Hypothesis Testing
- Data Snooping
- Non-Normal Distributions
- Multiple Testing Problem
- Model Assumptions
back testing
Detailed Mechanics of Backtesting
1.Data Collection
○ Sources
○ Types of Data
○ Data Cleaning
2. Strategy Definition
○ Trade Criteria:
○ Portfolio and Capital Allocation:
○ Risk Management Rules:
3. Implementation
○ Coding the Strategy
○ Simulation
4. Performance Analysis
○ Benchmarking
○ Risk Metrics
Expanded Key Metrics in Backtesting
- Annualized returns
- Sortini Ratio (like sharp but only consider downside volatility)
- Calmar ratio
- Percent Profitable Trades
- Profit Factor
- Omega Ratio
Benefits of Backtesting: A Deeper Insight
- Confidence Building
- Strategy Evolution
- Capital Allocation Guidance
Pitfalls and Considerations: Beyond the Basics
Parameter Sensitivity
Market Regimef Changes
Post- Strategy Drift
Adaptative Markets
Advanced Best Practices
Walk Forward Analysis
Monte Carlo Simulations
Stress Testing
Why Model Evaluation is Essential
Stakeholder Confidence
Strategic decision making
Regulatoru Compliance
Continuous Improvement
Evaluating Forecast Accuracy
Mean Absolute Error (MAE)
Mean Squared Error (MSE)
Root Mean Squared Error (RMSE)
Mean Absolute Percentage Error (MAPe)
Theil’s Statistic - compares the forecasted model to a (no-change) forecast
Classification Model Metrics (For Binary Outcomes, e.g., Default/No
Default)
- Confusion Matrix
- Accuracy
- Precision
- Recall (Sensitivity)
- F1 - Score
- Area under the ROC Curve
- Log Loss