Autoregressive conditional heteroskedasticity

Autoregressive conditional heteroskedasticity (ARCH) is like a video game where the rules change over time. In this game, the value of one thing depends on the values of the things that happened before it. This game also gets more - or less - difficult as you play. For example, if something bad happened in the past (like a stock market crash), the rules of the game change and the next turn could be more difficult than before. ARCH helps us understand how these changing rules affect what will happen in the future.