Study guides / CFA L1: CFA Program Level I / Statistics & probability
Statistics & probability
CFA L1: CFA Program Level I · quick review
Describing return distributions (center, dispersion, shape), probability rules, expected value, and portfolio math with covariance and correlation.
The rules
- Measures of center: mean, median, mode. Dispersion: range, variance, standard deviation, coefficient of variation (standard deviation ÷ mean).
- Skewness: positive (right) skew has a long right tail, with mean > median > mode; negative skew is the reverse. Kurtosis above 3 (excess kurtosis > 0) means fatter tails than normal.
- Probability: P(A or B) = P(A) + P(B) − P(A and B); P(A and B) = P(A | B) × P(B). Independent events: P(A | B) = P(A).
- Bayes' formula updates a probability with new information: P(A | B) = P(B | A) × P(A) ÷ P(B).
- Covariance shows how two variables move together; correlation = covariance ÷ (σ₁ × σ₂), between −1 and +1.
- Portfolio variance for two assets = w₁²σ₁² + w₂²σ₂² + 2w₁w₂ρσ₁σ₂. Lower correlation means more diversification.
- Normal distribution: about 68% within ±1σ, 95% within ±1.96σ, 99% within ±2.58σ. z-score = (x − μ) ÷ σ.
Formulas
- Coefficient of variation: CV = σ ÷ mean
- Correlation: ρ = Cov(A,B) ÷ (σA × σB)
- Two-asset portfolio variance: w₁²σ₁² + w₂²σ₂² + 2w₁w₂ρσ₁σ₂
- Bayes' formula: P(A|B) = P(B|A)P(A) ÷ P(B)
- z-score: z = (x − μ) ÷ σ
Common traps
- Correlation measures linear association only, and it doesn't imply causation.
- Lognormal distributions are bounded at zero and right-skewed, which suits asset prices, not returns.
- Excess kurtosis is kurtosis minus 3.
How it's tested
Interpreting skew and kurtosis, probability and Bayes calculations, and two-asset portfolio risk.
Practise it
The full lesson, worked examples and practice questions are in Tickmark, where every answer choice is explained.
Open the full lesson