arXiv++ Combinatorics

This paper is about the following question: How many riffle shuffles mix a deck of card for games such as blackjack and bridge? An object that comes up in answering this question is the descent polynomial associated with pairs of decks, where the decks are allowed to have repeated cards. We prove that the problem of computing the descent polynomial given a pair of decks is $#P$-complete. We also prove that the coefficients of these polynomials can be approximated using the bell curve. However, as must be expected in view of the $#P$-completeness result, approximations using the bell curve are not good enough to answer our question. Some of our answers to the main question are supported by theorems, and others are based on experiments supported by heuristic arguments. In the introduction, we carefully discuss the validity of our answers.

Normal approximations for descents and inversions of permutations of the set $\{1,2,...,n\}$ are well known. A number of sequences that occur in practice, such as the human genome and other genomes, contain many repeated elements. Motivated by such examples, we consider the number of inversions of a permutation $π(1), π(2),...,π(n)$ of a multiset with $n$ elements, which is the number of pairs $(i,j)$ with $1\leq i < j \leq n$ and $π(i)>π(j)$. The number of descents is the number of $i$ in the range $1\leq i < n$ such that $π(i) > π(i+1)$. We prove that, appropriately normalized, the distribution of both inversions and descents of a random permutation of the multiset approaches the normal distribution as $n\to\infty$, provided that the permutation is equally likely to be any possible permutation of the multiset and no element occurs more than $αn$ times in the multiset for a fixed $α$ with $0<α< 1$. Both normal approximation theorems are proved using the size biased version of Stein's method of auxiliary randomization and are accompanied by error bounds.

By a well-known result of Bayer and Diaconis, the maximum entropy model of the common riffle shuffle implies that the number of riffle shuffles necessary to mix a standard deck of 52 cards is either 7 or 11--with the former number applying when the metric used to define mixing is the total variation distance and the latter when it is the separation distance. This and other related results assume all 52 cards in the deck to be distinct and require all $52!$ permutations of the deck to be almost equally likely for the deck to be considered well mixed. In many instances, not all cards in the deck are distinct and only the sets of cards dealt out to players, and not the order in which they are dealt out to each player, needs to be random. We derive transition probabilities under riffle shuffles between decks with repeated cards to cover some instances of the type just described. We focus on decks with cards all of which are labeled either 1 or 2 and describe the consequences of having a symmetric starting deck of the form $1,...,1,2,...,2$ or $1,2,...,1,2$. Finally, we consider mixing times for common card games.