Mathematics of Card Shuffling
Statistics & Probability
How many shuffles does it take to randomize a deck of cards?
Introduction
The Gilbert-Shannon-Reeds model of a riffle shuffle shows that the total variation distance from a fully random deck drops sharply around 7 shuffles for a standard 52-card deck — the commonly quoted ‘seven shuffles’ figure is this cutoff, not an exact threshold. This exploration combines combinatorics, permutations and statistical tests: define a measure of randomness, track how quickly a real shuffle approaches it, and test different shuffle types (riffle, overhand, pile) against that measure.
Guiding Questions
- How can you define and measure 'randomness' mathematically for a shuffled deck?
- What is the probability that a specific card ends up in a specific position after n riffle shuffles?
- How does the shuffle type (riffle, overhand, pile) affect how quickly the deck approaches your randomness measure?
- Track a marked card through repeated real shuffles you perform yourself. How does its position after each shuffle compare with your theoretical model?
- Which real, physical shuffles could you perform and record yourself, so the data comes from your own experiment rather than a published one?