Statistical Foundations for Analyzing Trading Data
Summary
This introductory article connects probability and statistics to trading, using repeated coin flips to explain why larger samples tend to represent underlying probabilities more reliably. It frames a trader’s task as using observed trade or price data to estimate possible future behavior, while noting that risk assessment is outside its scope.
It defines sample mean, variance, skewness, kurtosis, covariance, and correlation, and illustrates them with height and weight data. It then links the sample mean to a moving average and describes how statistical measures can help characterize market series and relationships. MQL5 library functions are mentioned as a way to calculate these quantities. The examples are educational rather than market evidence: they use a small nonfinancial dataset and do not establish predictive power, trading profitability, or specific decision thresholds. Estimates from finite samples can be uncertain, and the article does not develop those limitations or cover risk management in depth.
Key ideas
- Probability describes possible outcomes, while statistics summarizes observed outcomes and can be used to estimate probabilities.
- Larger samples can provide a more representative estimate of an underlying process than very small samples.
- The sample mean measures a dataset’s center, while variance measures its spread.
- Skewness and kurtosis describe aspects of a distribution’s shape.
- Covariance and correlation summarize linear relationships, with correlation scaled to a bounded range.
- A moving average is an application of the arithmetic mean to a rolling window of prices.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.