Notes of CFA Level1 READING 8:STATISTICAL CONCEPTS AND MARKET RETURNS

kurtosis n. [统计学]峰态

lack of symmetry (skewness 偏斜)

测量尺度（scale of measure）或称度量水平（level of measurement）、度量类别，是统计学和定量研究中，对不同种类的数据，依据其尺度水平所划分的类别，这些尺度水平分别为：名目（nominal）、次序（ordinal）、等距（interval）、等比（ratio）。

harmonic mean [数]调和平均数

For values that are not all equal: harmonic mean < geometric mean < arithmetic mean. This mathematical fact is the basis for the claimed benefit of purchasing the same dollar amount of mutual fund shares each month or each week（定投）. Some refer to this practice as “dollar cost averaging.”

Quintile—the distribution is divided into fifths.
Decile—the distribution is divided into tenths.

The formula for the position of the observation at a given percentile, y, with n data points sorted in ascending order is：

Ly=（n+1）y/100

range = maximum value − minimum value

mean absolute deviation (MAD) 平均绝对偏差

population variance总体方差 & population standard deviation 总体标准偏差

sample variance 样本方差

Chebyshev’s inequality states that for any set of observations, whether sample or population data and regardless of the shape of the distribution, the percentage of the observations that lie within k standard deviations of the mean is at least 1 − 1/k2 for all k > 1.

According to Chebyshev’s inequality, the following relationships hold for any distribution. At least:

• 36% of observations lie within ±1.25 standard deviations of the mean.
• 56% of observations lie within ±1.50 standard deviations of the mean.
• 75% of observations lie within ±2 standard deviations of the mean.
• 89% of observations lie within ±3 standard deviations of the mean.
• 94% of observations lie within ±4 standard deviations of the mean.

coefficient of variation (CV）：CV=standard deviation/average value

In an investments setting, the CV is used to measure the risk (variability) per unit of expected return (mean).

The Sharpe Ratio(or reward-to-variability ratio) 收益对风险比率

Analysts should be aware of two limitations of the Sharpe ratio: (1) If two portfolios have negative Sharpe ratios, it is not necessarily true that the higher Sharpe ratio impliessuperior risk-adjusted performance. Increasing risk moves a negative Sharpe ratio closer to zero (i.e., higher). (2) The Sharpe ratio is useful when standard deviation is an appropriate measure of risk. However, investment strategies with option characteristics have asymmetric return distributions, reflecting a large probability of small gains coupled with a small probability of large losses. In such cases, standard deviation may underestimate risk and produce Sharpe ratios that are too high.

Outliers are observations with extraordinarily large values, either positive or negative.

A positively skewed distribution is characterized by many outliers in the upper region, or right tail. A positively skewed distribution is said to be skewed right because of its relatively long upper (right) tail.

A negatively skewed distribution has a disproportionately large amount of outliers that fall within its lower (left) tail. A negatively skewed distribution is said to be skewed left because of its long lower tail.

 

Kurtosis is a measure of the degree to which a distribution is more or less “peaked” than a normal distribution.

Kurtosis is a measure of the degree to which a distribution is more or less “peaked” than a normal distribution. Leptokurtic describes a distribution that is more peaked than a normal distribution, whereas platykurtic refers to a distribution that is less
peaked, or flatter than a normal distribution. A distribution is mesokurtic if it has the same kurtosis as a normal distribution.

Measures of Sample Skew /skju/ 偏斜 and Kurtosis /kɜː'təʊsɪs/峰态

Sample skewness is equal to the sum of the cubed deviations from the mean divided by the cubed standard deviation and by the number of observations.

s=sample standard deviation,样本标准差

When a distribution is right skewed, sample skewness is positive because the deviations above the mean are larger on
average. A left-skewed distribution has a negative sample skewness.

Sample kurtosis is measured using deviations raised to the fourth power.

excess kurtosis = sample kurtosis − 3

To interpret kurtosis, note that it is measured relative to the kurtosis of a normal distribution, which is 3. Positive values of excess kurtosis indicate a distribution that is leptokurtic (more peaked, fat tails), whereas negative values indicate a platykurtic distribution (less peaked, thin tails).