Internal Design of kvr2 | Overview | Data Flow Architecture | 1. Extraction Phase (values_lm) | 2. Computation and Metadata Injection (r2, comp_model) | 3. Dispatch and Visualization Phase | Handling Non-Standard Evaluation (NSE) | Dependency Strategy | Troubleshooting and Maintenance

Introduction: Why $R^2$ is Not Unique | The Eight + One Definitions | Standard Definitions | Definitions for No-Intercept Models | Robust Definition | When $R^2$ Goes Negative: Interpretation and Risks | Meaning of Negative Values | Case Study: Forcing a No-Intercept Model | The Transformation Trap (Power Models) | Distinguishing Variable Names from Functions | Transparency and Metadata: Beyond the Numbers | Inspecting Model Information | Enhanced Console Output | Technical Note: How R Calculates | The Shift in Baseline | Case Study: The Danger of No-Intercept Models | The Trap | Visualizing the Sensitivity of R-squared | Comparing Intercept vs. No-Intercept Models | Practical Example: The Sensitivity of $R^2$ | 1. The Inflation of $R^2_2$ | 2. The Drop in Predictive Accuracy | Adjusted R-squared Comparison | Visualizing the Comparison: The Diagnostic Dashboard | Key Features of the Dashboard: | Example: When $R^2$ Breaks (The Importance of Visuals) | A Note on Plot Customization | Conclusion: A Multi-Metric Approach | References