A data scientist tackles a probability challenge originally posed by popular educator Grant Sanderson of 3Blue1Brown, who creates acclaimed YouTube videos on mathematics. The puzzle involves randomly selecting string ends from a box and tying them together until no strings remain, asking how many loops form on average. With 50 strings as the target, brute force methods become unwieldy.

The elegant solution hinges on four key observations about the random process. Each round always eliminates one string regardless of outcome, making the total number of rounds deterministic. More importantly, rounds operate independently when counting loops, allowing the problem to decompose into individual expected values. For a single round with S strings, the probability of creating a loop equals 1/(2S-1), derived from the available remaining ends.

Summing this probability across all rounds from 1 to 50 strings yields approximately 2.94 expected loops. The author also implements a Monte Carlo simulation with 10,000 trials as validation, demonstrating how computational methods complement analytical solutions. This approach showcases how breaking complex problems into independent components often reveals simpler paths to solution.

The exercise illustrates why maintaining mathematical reasoning skills matters even in an AI-assisted era. By working through Sanderson's puzzle manually, readers practice decomposing probabilistic processes into manageable pieces—a skill that translates directly to real-world data science challenges where AI tools may not always apply.