Predicting Tomorrow: A Beginner’s Guide to AR and MA in Time Series

Niels Bohr once said that “prediction is very difficult, especially about the future”, and he wasn’t kidding! But why do we humans keep trying to peek into tomorrow? Well, decision-makers everywhere, from government officials (desperately trying to avoid the next crisis) to company executives (hoping not to become the next Zillow) need to anticipate what’s coming around the corner. In this article, I’ll walk you through the fascinating world of time series forecasting. We’ll explore two key concepts: Autoregressive (AR) models, which are like data fortune-tellers looking at past patterns, and Moving Average (MA) models, which smooth out the noise. I’ll break down how these models work, their intuitive foundations, and because nothing’s perfect, their limitations.