The mixing time of the Thorp shuffle

Ben Morris

DOI: https://doi.org/10.48550/arXiv.math/0507307

2005-07-15

Abstract:The Thorp shuffle is defined as follows. Cut the deck into two equal piles. Drop the first card from the left pile or the right pile according to the outcome of a fair coin flip; then drop from the other pile. Continue this way until both piles are empty. We show that the mixing time for the Thorp shuffle with $2^d$ cards is polynomial in $d$.

Probability

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to determine the mixing time of Thorp shuffle (a specific shuffling method). Specifically, the author Ben Morris studied the number of steps required for this shuffling method to fully randomize a deck of cards and proved that its mixing time is polynomial.

### Problem Background

1. **Definition of Thorp Shuffle**:
- Divide a deck of cards into two equal parts.
- Each time, take a card from the left or right pile according to the result of a fair coin and put it into a new pile until both piles are empty.

2. **Definition of Mixing Time**:
- Mixing time refers to the time required to make the card order sufficiently randomized through multiple shuffles.
- Mathematically, the mixing time can be formalized as the minimum number of steps such that the difference between the shuffled distribution and the uniform distribution is small enough.

3. **Status of Existing Research**:
- The upper and lower bounds of most natural shuffling methods have been well studied and solved.
- Due to its complexity, Thorp shuffle has been one of the unsolved problems.

### Main Contributions of the Paper

- **Giving the Polynomial - level Upper Bound for the First Time**: The author proved that the mixing time of Thorp shuffle is polynomial in terms of \(d\) (that is, the number of cards is \(2^d\)).
- **Using the Evolving Set Technique**: Introduced the concept of evolving sets to analyze the mixing time of Markov chains, which is a new technical means.
- **Combining the \(l_2\) Trick**: Used the method of combining the \(l_2\) norm and evolving sets to derive the upper bound of the root profile function.

### Main Results

- **Theorem 1**: The mixing time of Thorp shuffle is \(O(d^{44})\).
- **Bounds of the Root Profile Function**: Proved the upper bound of the mixing time by analyzing the root profile function.

### Method Overview

- **Evolving Set**: A technique for analyzing the mixing time of Markov chains, which can effectively handle complex shuffling processes.
- **\(l_2\) Trick**: Combined with \(l_2\) - norm analysis, provides a method for quantifying the change in probability distribution during the shuffling process.
- **Chameleon Process**: An extended shuffling model that simplifies the analysis by introducing color changes.

### Conclusion

This paper solves the long - standing problem of the mixing time of Thorp shuffle, provides the first polynomial - level upper bound, and introduces new analytical tools and techniques, laying the foundation for follow - up research.