The Top-to-Random Shuffle

The top-to-random shuffle is one of the simplest ways to mathematically model shuffling a deck: take the card currently on top and reinsert it at a position chosen uniformly at random among all $n$ spots (including possibly back on top). Repeat many times. It is far weaker than a real riffle shuffle in any single step, but its extreme simplicity makes it a favorite toy example for analyzing how a Markov chain converges to a uniform, fully mixed distribution over all $n!$ orderings of the deck.

The classic result, due to Aldous and Diaconis, is a coupon-collector-style argument: consider the original bottom card. Each shuffle moves one card to a random position, and the deck is fully "randomized" from the point of view of relative order roughly once every card that started below the original bottom card has, at some point, been moved to a position below it. This takes about $n \ln n$ shuffles to get close to uniform, with a sharp cutoff — well before $n \ln n$ shuffles the deck is still far from mixed, and shortly after, it is essentially fully random. For a standard 52-card deck, $n \ln n \approx 52 \times 3.95 \approx 205$ shuffles.

This "cutoff phenomenon" — a sudden transition from unmixed to mixed rather than a gradual fade — is the point of the example: it shows that "how many shuffles until random" can have a sharp, well-defined answer rather than being a matter of degree, a useful intuition for any Markov chain mixing-time question.