Mean, Median, Spread, Skew, and Fat Tails

Learn the five numbers that describe any list of returns: average, middle, spread, lopsidedness, and how often big moves happen.

Category: Statistics

In plain English

Imagine you write down one number every day, like how much a stock moved. After a while you have a long list. Instead of staring at the whole list, you boil it down to a few summary numbers: where the middle is, how spread out the numbers are, and whether one side sticks out more than the other. These summaries are the bread and butter of statistics, and they are how anyone makes sense of returns, prices, or any other column of data.

What it measures

Five summary numbers do most of the work. The mean is the plain average, and the median is the middle value once you sort the list. Variance, and its square root the standard deviation, measure spread, meaning how far the numbers typically sit from the average. Skewness says whether the list is lopsided, and kurtosis says whether extreme values happen more often than a smooth bell-shaped curve would predict (this is what people mean by fat tails). (This small example uses the population standard deviation; software usually uses the sample version, dividing by n-1, which gives about 1.0% here — either way the numbers cluster near the average.) - Mean: add 2, 4, 9 and divide by 3 = 5. The median (middle of sorted 2, 4, 9) = 4. - Median is safer when one giant value distorts things: for 2, 4, 90 the mean jumps to 32, but the median stays 4. - Standard deviation: for daily moves of -1%, 0%, +1%, the spread is about 0.8%, so the numbers cluster near the average. - Skewness above 0 means a long tail on the high side; below 0 means a long tail on the low side; 0 is symmetric. - Kurtosis above 3 (the bell-curve value) means fat tails: big up or down days show up more often than a bell curve says.

How SurgeFlow uses it

These five descriptions sit underneath almost every return and volatility number on the site. When a metric reports an average return it is using the mean; when it reports volatility it is using standard deviation. The Factors and Backtesting tabs lean on these to summarize how a strategy behaved over time, and any chart of daily returns is really a picture of this spread and shape. SurgeFlow uses them to describe what the data did, never to predict what it will do next. - Mean and median: typical return of a stock, factor, or strategy. - Standard deviation: the volatility number you see next to returns. - Skewness and kurtosis: warnings that returns are lopsided or prone to big jumps.

How to read it on the site

Open the Backtesting tab (/backtest/benchmark) or the Factors tab (/factors) and look at the summary stats next to any return series. A higher standard deviation means a bumpier ride, not necessarily a better or worse one. If the mean is far from the median, treat the mean with caution, because a few extreme days are pulling it. If kurtosis is high, expect the occasional shock day that a calm-looking average hides.

What it cannot tell you

These numbers describe the past list of values; they are not a forecast of the future. A low standard deviation does not promise safety, and a high average return measured over a short window can be mostly luck. Fat tails warn that rare big moves happen, but they cannot tell you when one will arrive or how large it will be. Always remember that a summary throws information away on purpose, and SurgeFlow is research and education, not investment advice. - Past spread and shape do not predict future spread and shape. - A single outlier can quietly distort the mean, so check the median too. - High kurtosis flags that shocks are possible, not when they will hit. - These are descriptive tools; they give no buy or sell signal.