Beginner Level
What Is It?
A z-score (standard score) measures how many standard deviations a data point is from the mean. It standardizes values, enabling comparison across different scales and distributions.
Origin
Z-scores emerged from the standardization of normal distributions in the early 20th century. Fisher and others used them for statistical tables. Now ubiquitous across statistics, finance, and quality control.
Why It Matters
Z-scores enable meaningful comparison of values from different distributions. They identify outliers and extreme events. Understanding z-scores is essential for probability calculations, hypothesis testing, and risk assessment.
Intermediate Level
Market Mechanics
Calculation: z = (x - μ) / σ. Positive z: above average; negative z: below average. |z| > 2: approximately 5% of normal distribution; |z| > 3: approximately 0.3%. Standard normal table maps z to probability. Z-scores assume normal distribution—real returns have fatter tails.
How It Behaves
Market returns show more extreme z-scores than normal distribution predicts (fat tails). Z-scores cluster during crises. Mean reversion may follow extreme z-scores. Different assets have different z-score distributions. Bollinger Bands use z-score logic.
Key Data to Watch
- Z-scores for returns and indicators
- Percentage of observations beyond |z| > 2 or 3
- Maximum z-score observed
- Z-score distributions by asset
- Extreme z-score clustering
- Fat tail measures
Advanced Level
Institutional Behavior
Quants use z-scores for normalization and signal generation. Risk managers flag extreme z-scores for investigation. Quality control uses z-scores for process monitoring. Stress testing includes extreme z-scenarios. Machine learning normalizes using z-scores.
Professional Use Cases
- Outlier detection
- Signal normalization
- Probability estimation
- Process control
- Stress scenario design
- Feature scaling for ML
AI Interpretation in Systems Like Arkhe
- Normalization Agent: Converts indicators to z-scores for comparison
- Risk Agent: Flags extreme z-scores as potential risks
- Signal Agent: Generates signals based on z-score extremes
Key Takeaways
Z-scores standardize values for comparison and analysis. While assuming normality, they remain useful even with fat-tailed distributions. Understanding z-scores enables better probability assessment and risk management.