Magical

Applications of Mathematics

Book Review of Magical Mathematics: The Mathematical Ideas That Animate Great
Magic Tricks, by Persi Diaconis and Ron Graham. Reviewed by Arthur Benjamin.

Magical Mathematics exemplifies how mathematics should be taught. Imagine that
you begin your class in discrete mathematics (or number theory or mathematical
modeling) with the following trick. You bring out a deck of cards, and ask someone
in the audience to cut the cards as often as they like, then take the card on top of the
deck. Pass the deck to someone else have them take the next card, and have a third
person take the following card. Ask each person to concentrate on their card. You
explain that you will determine one persons suit with just a few hints. The first
person tells you their value (Ace or two or or King) and the second person tells
you their suit (Spades or Hearts or Diamonds or Clubs), and the third person tells
you nothing. With that information, and a little extra concentration, you reveal
everyones card. Most likely everyone in the class is now wondering How did you
do that? This is precisely the reaction that we strive for in our classrooms.

Lets think about this trick. Since the cards were cut, but not shuffled, the cards are
still in the same cyclic order. Since there are 13 x 4 = 52 possible responses from the
first two audience members, that should be enough information to reveal the three
cards. If the cards were stacked in an obvious order (say beginning with Ace of
Spades, Two of Spades, ) and you heard that the first card had value Seven and the
second card was Spades then you know the identities of all three cards (Seven of
Spades, Eight of Spades and Nine of Spades). But of course, this ordering of the deck
would be obvious to your audience once they hear you name the cards (or if they
glance the deck while the cards are being cut). Is there a less obvious way to arrange
the cards? Heres a very effective stack (known to magicians as Si Stebbins). The
card values are arranged starting with the King (with numerical value thirteen) and
adding 3 (mod 13) to each value producing the order (K, 3, 6, 9, Q, 2, 5, 8, J, A, 4, 7,
10). The suits are arranged in CHaSeD order (Clubs, Hearts, Spades, Diamonds) so
the deck begins (KC, 3H, 6S, 9D, QC, ). You could casually show the cards before the
trick begins and they will appear to be in random order. Now if the first audience
member says 5 and the second member says hearts you can combine those
pieces of information to determine that the three cards must be 5 of Clubs, 8 of
Hearts, and Jack of Spades, in that order.

One could easily imagine going off on tangents to explore why adding 3 (mod 13)
results in cycles of length 13, and that adding any other constant would work just as
well because 13 is prime. The trick described here is just one of many from the
books chapter on Universal Cycles. More subtle variations are provided where you
ask five people to take a card like before, and ask everyone to concentrate. Explain
that you are getting mixed signals. Who has the largest valued card? (They compare
cards and someone raises their hand.) Then ask who has the second largest, then the
third largest card? Once these people are identified, you tell everyone what their
card is. Again, this requires that the cards are ordered in such a way that the

answers to your innocent sounding questions uniquely determines the card chosen
by the first (and subsequent) person. Other tricks are presented that utilize de
Bruijn sequences: binary sequences of length 2^k, where every string of length k
appears exactly once when the string is written in a circle. The authors generalize
these ideas (for example to higher dimensions), and offer applications, most of
which lead to more magical effects.

Although Magical Mathematics is written for a general audience, and both authors
have considerable experience as performers, they are first and foremost
mathematicians of the highest caliber, so their exposition should be especially
enjoyed by readers of this journal. Many of the tricks described in this book require
some thinking on the part of the performer, requiring more sleight of mind than
sleight of hand, but some of the tricks are self-working. Many of the tricks are
surprising, even after you know the secret.

My favorite chapter dealt with the Gilbreath Principle and its extensions. The
development of this concept and its generalizations and applications have exactly
the same appeal as watching a beautiful number pattern turn into a complex and
satisfying theorem. Heres Gilbreaths first principle in action. Suppose you have a
deck of cards where the cards are secretly arranged so that the colors alternate
(black, red, black, red, ). Ask your volunteer to deal about half of these cards into a
pile, then riffle shuffle the two halves together. Surprisingly, even after one riffle
shuffle, the cards still retain an interesting structure: the top two cards are still
black and red (in some order); the next two cards are still black and red; the next
two cards are still black and red; and so on. Not obvious, right?

How about this? Give your spectator 25 (secretly pre-arranged) cards. First have the
spectator demonstrate that they can deal cards, by dealing about half the cards into
second pile, then riffle shuffle both halves together. The spectator then deals five 5-
card poker hands, and everything seems normal. Then the spectator deals five new
5-card poker hands, only to see that one of the players was dealt a straight, and
another player was dealt a flush.


A Gilbreath shuffle on n cards consists of choosing a number j between 0 and n,
dealing j cards into one pile, then riffle shuffling the two piles together. Among the n!
possible permutations of n cards, how many of them are achievable after one
Gilbreath shuffle? Remarkably, the answer is 2n-1. The material is presented in such a
way that the reader (especially one who is already interested in mathematics) will
be just as interested in the proof of that statement as understanding how a magic
trick based on that principle works. The general Gilbreath principle says that if the
numbers 1 through n are given a Gilbreath shuffle, then for every integer j, the top j
numbers (and the next j numbers and the next j numbers, ) are all distinct mod j.
These theorems are then applied to create truly magical moments that will appeal to
many an audience, especially those sitting in a discrete mathematics classroom. If
that werent exciting enough, the authors then go on to show how this Gilbreath

shuffle is intimately connected to real periodic points of the Mandelbrot set. The fact
that there is any connection at all is amazing as any magic trick.

The Gilbreath shuffle is imperfect since there is no way to predict the exact order
of cards after one shuffle. But other shuffles, such as perfect shuffles (also known as
Faro shuffles), are not random at all, since they correspond to an exact permutation.
In a perfect shuffle, the cards are cut exactly in half, then the two halves are
interleaved perfectly. There are two types of perfect shuffles: an Outshuffle keeps
the top card on top; an Inshuffle moves the top card to the second position. The
authors describe ways to control cards using these shuffles. For example, to move
the top card to the nth position, you simply express n-1 in binary and follow
instructions. For instance, to bring the top card to position 42, we see that 41 = 32 +
8 + 1 = (101001)2 and then perform In-Out-In-Out-Out-In. (Even if you cant
perform a perfect shuffle, the result is beautiful!) There are still many open
questions about these shuffles. It was only recently discovered how to easily invert
the process to bring a card from a desired position to the top using perfect shuffles.

Another shuffle with attractive mathematical properties is the Down and Under
shuffle (or Australian shuffle) which magicians sometimes use to reveal a chosen
card. From a packet of cards, the magician deals the top card to the table, then one
card to the bottom of the packet, then one card to the table, then one card to the
bottom of the packet, and continues this until only one card remains in the hand.
The elegant mathematical theorem is that if the packet contains n cards, then and if
n = 2k + x, where 1 x n/2, then the last card to be dealt is the card originally in
position 2x.

The book provides an interesting history of mathematical magic, which has been in
existence for a very long time. Some tricks are described in one of the very first
books published on magic in 1584, as well as one of the first books on arithmetic
(Fibonaccis Liber Abaci) published in 1202. The authors even go back thousands of
years to discuss some of the mathematics that arises in the ancient Chinese text, the
I Ching, and suggest some performance ideas based on them.

The authors devote a chapter to the mathematics of juggling. Both magic and
juggling have a long history, dating back thousands of years. As the authors point
out, There is also a strong connection between mathematics and juggling.
Mathematics is often described as the science of patterns. Juggling can be thought of
as the art of controlling patterns in time and space.

Most of the mathematics and tricks in the book have a combinatorial or number
theoretic flavor, but other branches of mathematics are covered. There is a
clever topological swindle involving a simple loop of chain, as well as some discrete
probability. (The authors are still in search of a magic trick based on calculus.) Here
is a random trick that surprised me. Imagine you have 5 cards, two of which say Win
and three of which say Lose. The spectator shuffles the cards and they will be turned
up one at a time. You explain that you will place bets as follows. Each time you bet,

your next bet will be half as much if you win, and will be 1.5 times as much if you
lose. (For instance, if you place a $10 bet and win, your next bet is $5 if you win and
its $15 if you lose.) Starting with a bet of $16, what will be your profit or loss after
your 5 bets? Surprisingly, regardless of how the cards are shuffled, you end up with
a $5 profit. I can see this being turned into an interesting classroom exercise. Ill
save that exercise (and the elegant generalization) to the reader.

Towards the end of the book, we meet some of the most important contributors to
mathematical magic. Diaconis describes his visit to the home of Stewart James, a
reclusive genius who has made vast contributions to card magic and mathematical
magic. When Diaconis asked James if he could borrow a deck of cards, he was
shocked to learn that James had not had a deck of cards in his house for years, even
though he wrote a monthly column on card magic. James explained, This may
sound strange but I dont have a real deck of cards in the house havent had one for
four or five yearsAfter all, when Agatha Christie writes a murder mystery, she
doesnt have to go out and kill somebody. The authors show special reverence to
Martin Gardner, who wrote the foreword to this book, just before he passed away in
2010. Their style of writing is very reminiscent of Gardners, with most of the
material presented at a high school mathematics level, and with lots of examples to
make the exposition go down smoothly.

The book, published by Princeton University Press, is beautifully illustrated, printed
on high quality paper and loaded with color photos. Both authors have many great
stories and anecdotes to tell, but I have one small quibble. In relating a story the
authors most often used the pronoun we or us to describe a situation that clearly
only happened to one of them (e.g., When the time came for us to apply to graduate
school, Martin Gardner was one of our letter writers.). Although clearly the authors
have had many joint experiences together, it would have been less distracting for
there to be more use of expressions like one of us (P.D.) when it applied.

Mathematical Magic is a truly magical book, containing ample amounts of
mathematics and magic that will amaze and amuse. Diaconis and Graham are both
first-rate mathematicians and performers and offer insights and ideas that could not
have been expressed by anyone else. This book is destined to be a classic on the
subject.