Netflix and Disney+ seem to know what I want to watch before I do sometimes, which is impressive but also a bit creepy. I know it's maths-based, but is it just tracking what I've watched, or is there something more complex going on with prediction algorithms?
It's way more complex than just tracking what you've watched. These services use collaborative filtering (basically finding people similar to you and recommending what they liked), content-based filtering (analyzing the actual shows you watched to find similar ones), and machine learning models that look at tons of data points like how long you pause, rewind, or stop watching something, what time of day you watch, even your browsing patterns. They're also constantly testing different recommendations to see what keeps you engaged longest, which is why the algorithm gets spookily accurate over time - it's literally designed to predict your behavior because that's how they keep you subscribed.
Log into answer.
