Statistics & QR
Szymon Kopyciński · 23 September 2026
Statistics and quantitative research
Quant
Short for quantitative.
Depending on context, a quant may be a quantitative researcher, trader, developer, analyst or another mathematically oriented role.
Model
A simplified mathematical or statistical representation of some aspect of reality.
Feature
An input variable used by a model.
Examples include recent returns, volatility, order book imbalance and macroeconomic data.
Target
The quantity a model is trying to predict.
For example, the return over the next five minutes.
Observation
One data point or sample in a dataset.
Time series
Data indexed through time.
Asset prices and returns are classic examples.
Cross-sectional data
Data describing many different entities at one point or period in time.
For example, comparing valuation ratios across 500 stocks today.
Mean
The arithmetic average.
Median
The middle value after observations are ordered.
Variance
A measure of how dispersed values are around their mean.
Standard deviation
The square root of variance.
It is the standard measure of volatility.
Covariance
A measure of how two variables move together.
Correlation
A normalised measure of how strongly two variables move together.
Correlation ranges from −1 to +1.
Autocorrelation
Correlation between a time series and lagged versions of itself.
For example, whether today's return is related to yesterday's return.
Regression
A statistical method for modelling the relationship between variables.
A simple linear regression is written:
$$ y = \alpha + \beta x + \epsilon $$
Alpha
In regression, the intercept term.
In finance, alpha also means excess return or trading edge (see Alpha in section 7).
Beta
A coefficient describing sensitivity to another variable.
In asset pricing, market beta describes how strongly an asset moves relative to the market.
Residual
The difference between a model's prediction and the observed value.
Factor
A variable used to explain common variation in asset returns.
Examples include market, value, momentum, size and sector exposure.
Stationarity
The property that a time series' statistical properties, such as its mean and variance, remain stable over time.
Many statistical methods assume stationarity. Prices are typically non-stationary, whereas returns are much closer to stationary, which is one reason research is usually done on returns.
Distribution
A mathematical description of the possible values a random variable can take and how likely each is.
Normal distribution
The familiar bell-shaped probability distribution.
Simple models often assume normally distributed returns, but real returns have heavier tails and other deviations.
Fat tails / Heavy tails
A distribution with extreme outcomes more frequent than a normal distribution would predict.
Financial returns typically display fat tails.
Skewness
A measure of asymmetry in a distribution.
Kurtosis
A measure of the tail heaviness of a distribution.
Financial returns typically show excess kurtosis relative to a normal distribution.
Expected value
The probability-weighted average outcome of a random variable.
A strategy needs positive expected value after costs to be profitable over time.
Hypothesis test
A statistical procedure for assessing evidence against a specified null hypothesis.
P-value
The probability, assuming the null hypothesis is true, of observing a result at least as extreme as the one actually observed.
It is not the probability that the null hypothesis is true, nor the probability that your hypothesis is correct.
Statistical significance
A statement that an observed result passes a predefined statistical threshold.
Statistical significance does not imply economic significance or tradability: a statistically significant effect may be too small to survive costs.
Part of Glossary Index · Previous: Positions, PnL & Risk · Next: Backtesting & Validation
NEFS Quant · About · Blog · News & Events · People · Privacy policy