Part 10/16:
They modeled the web as a huge network of pages linking to each other, akin to a giant interconnected chain. The "importance" or PageRank of a page was determined by the probability that a random "surfer" clicking links would land on it. To avoid getting stuck in loops and ensure fair exploration, they introduced a "damping factor" representing random jumps—similar to Markov chain models with occasional teleportation to any page.
This mathematical model vastly improved search quality, ensuring the most relevant and authoritative pages rose to the top. When Google launched in 1998, their use of Markov chain-based PageRank catapulted them past naïve keyword-matching engines like Yahoo, which relied on simple frequency counts.