This article was downloaded by: [University of Arizona]

On: 29 July 2012, At: 03:33

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

BSHM Bulletin: Journal of the British

Society for the History of Mathematics
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/tbsh20

On mathematical games
a
Jorge Nuno Silva
a
University of Lisbon, Portugal

Version of record first published: 27 May 2011

To cite this article: Jorge Nuno Silva (2011): On mathematical games, BSHM Bulletin: Journal of the
British Society for the History of Mathematics, 26:2, 80-104

To link to this article: http://dx.doi.org/10.1080/17498430.2011.560511

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-

conditions

This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation
that the contents will be complete or accurate or up to date. The accuracy of any
instructions, formulae, and drug doses should be independently verified with primary
sources. The publisher shall not be liable for any loss, actions, claims, proceedings,
demand, or costs or damages whatsoever or howsoever caused arising directly or
indirectly in connection with or arising out of the use of this material.
 BSHM Bulletin
Volume 26 (2011) 80–104

On mathematical games
Jorge Nuno Silva
University of Lisbon, Portugal

In this paper we survey most of the games that, throughout history, can be classified as
mathematical games. This paper is based on a talk given at the conference ‘Numeracy:
historical, philosophical, and educational perspectives’ at St Anne’s College, Oxford,
December 2009.

To Galina, my beloved wife

Downloaded by [University of Arizona] at 03:33 29 July 2012

W
hat are mathematical games? How can we say that one board game is
more mathematical than another? The answers to these questions are not
easy to find. By mathematical games we mean more than the games
covered in the bible of combinatorial games (Berlekamp et al. 2001). We intend to
include board games whose practice resembles, in one way or another, a
mathematical way of thinking. It is hard to say how mathematical a game is,
however, it is not difficult to agree that chess and Go are more mathematical than
Monopoly. We therefore survey games that show some kinship with mathematics.
These connections are diverse and can be obvious or lie hidden in the deep rule
structure of the game. Our presentation unfolds essentially in chronological order,
from Babylon to nowadays.
Games of chance, like the very old dice and astragals (Figure 1), are ruled by the
mathematics of random processes. They have been around for several millennia.

Figure 1. Dice and Astragals

The Chinese game Wei-Qi, also known as Go (Figure 2), is extremely old.
Some say, four thousand years old. Played on the intersections of a 19  19 lattice, its
main rules are very simple and natural. As the former chess world champion Lasker
put it, if there is life in other planets, those beings will not know anything about
bishops, knights, kings, and queens of chess, but they will play Go for sure!
However, it has resisted mathematical analysis. Some of the most relevant

BSHM Bulletin ISSN 1749–8430 print/ISSN 1749–8341 online ß 2011 British Society for the History of Mathematics
http://www.tandf.co.uk/journals
DOI: 10.1080/17498430.2011.560511
 Volume 26 (2011) 81

mathematical studies of Go have been made by Berlekamp’s team in UC Berkeley

(Berlekamp and Wolfe 1994).
Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 2. Board and pieces of Go

The family of games that go by the name of Mancala (Figure 3) includes

hundreds of variants, mainly from Africa and Asia. Its origins are lost in time. The
players usually dig little holes on the ground and use seeds for pieces. Some can be
extremely complex. The nature of these games is clearly arithmetical, but they are
loaded with cultural and psychological relevance (De Voogt 1997).

Figure 3. Mancala board and seeds from Cameroon (author’s collection)

Archimedes created a puzzle with fourteen pieces, the Stomachion, that resembles
the Tangram (Figures 4 and 5). Tangram is essentially a solitaire toy (Slocum 2003)
but Stomachion is much more than that. We do not know exactly how Archimedes
used his puzzle, but it was essentially mathematical. Could it be that he was exploring
combinatorics? For instance, the calculation of the number of ways the pieces can be
arranged in a square proved to be a difficult problem that Bill Cutler solved in 2003
(Pegg 2003). The dramatic history of the extant text where Archimedes refers to it
can be found in the book by Netz and Noel (2007).
 82 BSHM Bulletin
Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 4. Tangram

Figure 5. Stomachion

The Arabs brought Alquerque (Figure 6) to Europe, which became very popular.
This forefather of Checkers and Draughts used a lined board and the action
happened on the intersections. Each player in turn could move a piece to an adjacent
empty intersection, along one of the marked lines. Captures used the small jump
method and a player could take multiple jumps and captures. The first capture was
mandatory. A player wins a game if he captures all his opponent’s pieces, or blocks
all his movements. If none of these situations looks attainable, the winner is the
player with more pieces on the board. There is a software program, Chinook, that has
beaten any human at Checkers and Draughts since 1996, and the game was
completely analysed and proved to be a draw, under perfect play, in 2007 (Schaeffer
 Volume 26 (2011) 83

Figure 6. Alquerque board and pieces in starting position

Downloaded by [University of Arizona] at 03:33 29 July 2012

et al. 2007). Games of Dice and Alquerque are treated in the main source of medieval
games: a magnificent book that Alfonso X commissioned in 1283 (Figure 7).

Figure 7. Dice and Alquerque in Alfonso’s Los libros de acedrex

Chess was also brought by the Arabs from the East to Europe (Figure 8).
Originally, this game represented a battle with its soldiers, elephants, and men on
horses. In Europe the names and shapes of the pieces evolved and eventually
emulated the medieval society, with Kings, Queens, Bishops, Knights, and
peasants (Murray 1913; Lhôte 1993). The connections between chess and
mathematics have been stressed (Petkovich 1997; Watkins 2007). As Hardy
(1992) put it, ‘a chess problem is pure mathematics’. It has been noticed with
surprise that we can find wunderkinder in only three intellectual areas:
mathematics, music, and chess.
The rules of the game make it natural to define the distance between two squares
of the board as the minimum number of moves it takes to a King to go from one to
the other. There are several paths that attain this minimum, some of them looking
very ‘long’. The strange properties of the corresponding topology find multiple
tactical and strategic applications in real play. There are some famous puzzles arising
from chess. Leonhard Euler (1766) studied the Knight tour (Figure 9), where the task
consists in finding the right sequence of moves that leads a Knight to visit every
square of the board just once.
 84 BSHM Bulletin
Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 8. Illustration of a chess game in Alfonso X’s Los libros de acedrex

Figure 9. An example from Euler’s original work

The game of Backgammon (Figure 10), very popular today, has its origins in
Ancient Babylon. At the beginning of the twentieth century, close to Ur,
archaeological field explorations unearthed several sets of boards and pieces of a
magnificent game, which apparently had been buried with their wealthy owners
(Murray 1951). This game, played some 4500 years ago, is essentially a race, the
winner being the player who first finishes the trail with his pieces. The dice had a
particular shape: they were triangular pyramids. The exact rules of this game were
found in two clay tablets and translated by I Finkel. Thus, this is the oldest game we
can play according to its original rules.
The Game of Thirty Squares, or Senet (Figure 11), was popular in Egypt 2500
years ago. Its rules are believed to be different from the Royal Game of Ur, but the
 Volume 26 (2011) 85

Figure 10. The Royal Game of Ur

Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 11. A modern Senet board

nature of the game is the same: a race. The British Museum exhibits well-preserved
exemplars of both games.
In Rome, Duodecima Scripta (Figure 12) followed the same principles of a race
between two teams. In the Middle Ages, the game of Tables looked already like
our Backgammon. There were several variants of Tables as well (Figure 13). These
games used, besides the players’ tokens, cubic dice. The rules are based on the

Figure 12. Schematic board of Duodecima Scripta

 86 BSHM Bulletin
Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 13. A variant of Tables in Alfonso’s Los libros de acedrex

movement of the pieces according to the throws of the dice. Pieces that lie alone in
one position can be thrown out of the board and have to start the trail from the
beginning again.
The games mentioned above were popular among all classes of people, rich and
poor, educated and illiterate. Not all the games had such broad habitats. One
example is Ludus Regularis. By the end of the tenth century, as dice games were
banned by the Church, the priest Wibold, in Noyon, South of France, created a
thematic dice game with a virtuous theme. Ludus Regularis’s description reached us
due to fortunate circumstances (Le Glay 1834). This game used three cubical dice
with vowels in the place of the usual dots, and a tetrahedral die with consonants
(Figure 14). As there are 56 ways of throwing the cubic dice, Wibold listed 56 virtues.
His list starts:

1:1:1: Karitas:3:
1:1:2: Fides:4:
1:1:3: Spes:5:
1:1:4: Justitia:6:
1:1:5: Prudentia7:
1:1:6: Temperantia:8:
1:2:2: Fortitudo:5:
1:2:3: Pax:6:
1:2:4: Castitas:7:
1:2:5: Misericordia:8:
1:2:6: Obedientia:9:
1:3:3: Timor:7:
 Volume 26 (2011) 87
Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 14. A page of Le Glay

Throwing the dice the players tried to match numbers and letters of virtues. If they
did, those virtues would become theirs. Two virtues with numbers adding to 21 were
considered linked and some special rules applied to them. It is worth pointing out
that these two numbers, 56 and 21, appear in several other games throughout history
(for example, in Tarot there are 56 minor arcana and 21 numbered major arcana).
Wibold elaborates on his choices of numbers for the virtues. His approach is
essentially Pythagorean.
The nature of the next game of our survey, however, is deeply linked with the
mathematical tradition of old medieval Europe, which brings us to its most
influential author, Boethius. Anicius Manlius Severinus Boethius was born in Rome,
around 480, in a powerful family, the Anicii. His father, a successful politician, died
early, leaving Boethius to be educated by Aurelius Symmachus, an expert in the
 88 BSHM Bulletin

Greek classics, who inspired his pupil to follow in his footsteps. The latter was a
bright student, reaching, at young age, a superior level in the seven liberal arts: logic,
rhetoric, grammar, arithmetic, music, geometry, and astronomy. Arithmetic (Boèce
1995) was one of his own works (mainly a translation from Nicomachus of Gerasa
(Nicomachus 1952)), as well as a book on geometry. They showed his understanding
of the liberal arts as preparation for studying philosophy. His works were in the
academic curricula for centuries in Europe, and some of his texts were still used in
the eighteenth century in several European universities. His political career started
well, being a consul like his father, but he fell victim of Theodoric’s wrath and was
executed around 525 AD.
In the introduction to Arithmetic, which he dedicated to Symmachus, Boethius
explained his intention of publishing a text on each of the four components of the
quadrivium: arithmetic, music, geometry, and astronomy. These made up what
Boethius called the fourfold path to philosophy. In this mathematical tradition we
Downloaded by [University of Arizona] at 03:33 29 July 2012

don’t find any structured explanations as we do in Euclid. The arithmetical results

are unproved and the mystic characteristics of numbers are emphasized.
Arithmetic contains mainly definitions of kinds of numbers (even, odd, prime,
composite, perfect, and so on) and their relations. His classification of the
relations between two numbers was based on ratios. Thus, two numbers can be in
multiplex ratio (like n:1), superparticular ((n þ 1) : n), or superpartient ((n þ m):n, for
some m > 1).
Boethius also considered means of two numbers, namely the arithmetic,
geometric, and harmonic means. Besides showing their relations to music,
Boethius claimed that their major virtue resides in guiding the students’ mind to
God, highlighting the unity of arithmetic, philosophy, and theology. Boethius
mentioned that the ancient scholars, of whom Pythagoras was the most important
reference, used a quadruple way to philosophical knowledge. It consisted of:
arithmetic (the study of numbers in themselves), music (the study of the relations
between numbers), geometry (the study of magnitudes), and astronomy (the study of
moving magnitudes). Arithmetic is more than logistics and geometry lies above
geodesics. Number has an independent existence, the mathematician just uncovers
and contemplates it. The nature of these ideal entities, mathematical objects, gives
them a superior rank.
Then came the classification of the numbers into several categories, as even–
odd, prime–composite, evenly–even (powers of 2), and so on, along with
characterizations of these classes. The Sieve of Erathostenes is presented, as well
as the Euclidean Algorithm for the greatest common divisor. Consistently, no
proofs are provided. Like Nicomachus, Boethius defined perfect numbers and gave
an expression to generate them. Perfect numbers are those that equal the sum of
their proper divisors, like 6 ¼ 1 þ 2 þ 3. In Euclid’s Elements we can find a proof
that the formula (in modern notation) n ¼ 2p1 (2p  1) where 2p  1 is prime,
generates perfect numbers. Euler showed that all the even perfect numbers are
given by this formula (we still do not know whether there are any odd perfect
numbers). Boethius, like Nicomachus before him, erred when he claimed that the
perfect numbers alternate in the last digit between 6 and 8. (The perfect numbers
known in antiquity were 6, 28, 496, 8128.)
Next there was an introduction of Pythagorean character, identifying
equality with Same and inequality with Other, which refers to Plato’s Timeaus
(Cornford 1997). Then, after elaborating on several kinds of multiples, Boethius
 Volume 26 (2011) 89

stated a principle that is quoted in most manuals of Rithmomachia:

‘Every inequality comes from equality’. The process that demonstrates this
consists of constructing a matrix aij, with three columns. Successive rows are
obtained according to the following rules an1 ¼ a(n1)1, an2 ¼ a(n1)1 þ a(n1)2,
an3 ¼ a(n1)1 þ 2a(n1)2 þ a(n1)3.
As an example Boethius offered the following, where a11 ¼ a12 ¼ a13 ¼ 1:

1 1 1
1 2 4
1 3 9
  
  
  
Downloaded by [University of Arizona] at 03:33 29 July 2012

From the first row, which consists of a repeated number, we get the sequence of the
doubles, then the one of the triples, and so on. Clearly, all multiples will eventually
appear in these lists. Boethius’ reverse method can be described as follows. Start with
three numbers in geometrical progression, a, ra, r2a, and construct an (r  1)  3
matrix (aij) in which the first row is the given progression and the following rows are
given by the rules: an1 ¼ a(n1)1, an2 ¼ a(n1)2  a(n1)1, an3 ¼ a(n1)3  a(n1)1
 2(a(n1)2  a(n1)1). Accordingly, we get

a ra r2 a
a ðr  1Þa ðr  1Þ2 a
a ðr  2Þa ðr  2Þ2 a
  
  
  

Thus, in the rth row we will reach equality.

Boethius exemplified this starting with the first row 8, 32, 128, which corresponds
to the parameters a ¼ 8, r ¼ 4:
8 32 128
8 24 72
8 16 32
8 8 8

Boethius later began the study of the numbers in themselves, not in relation to
each other. In his opinion, aside from the implicit interest, this knowledge would be
useful when studying progressions. Numbers can be linear, plane, and solid, their
nature arising from their geometrical figurate representation. Unity, 1, has no
dimensions. Linear numbers start with 2 and are those that can be obtained by
successively adding units in a line:

jj jjj jjjj jjjjj jjjjjj jjjjjjj jjjjjjjj

Triangular numbers are the first plane numbers to appear. According to

Boethius, the first triangular number is 3, and then come the numbers that lie evenly
 90 BSHM Bulletin

along three equal sides.

Downloaded by [University of Arizona] at 03:33 29 July 2012

The generation rule for these numbers is given. In symbolic terms, we have
Tn ¼ 1 þ    þ n
(Boethius considered 1 as a particular case of a triangular number).
Square numbers are those that extend themselves within four angles keeping
equal sides.

Boethius shows a procedure to obtain the square numbers from the odd numbers:
Qn ¼ 1 þ    þ 2n  1:

Next came the pentagonal, hexagonal, heptagonal numbers. Boethius gave

relations between these families and emphasized the role of the triangular numbers:
Qn ¼ Tn þ Tn1 ;
Pn ¼ Sn þ Tn1 ;
Hexn ¼ Pn þ Tn ;
Hepn ¼ Hexn þ Tn1 :
He also referred to oblong numbers (On ¼ n(n þ 1)) noting that, just as the odd
numbers generate the squares, the even numbers generate the oblongs, given that
On ¼ 2 þ    þ 2n:
 Volume 26 (2011) 91

For Boethius, everything was a mix of Same and Other, the two Platonic principles.
Odd and even numbers play those roles in the world of numbers.
The study of proportions was advised because of its utility in music,
astronomy, and geometry, besides being of great help in better understanding
of the classics. According to Boethius, Plato, Aristotle, and Pythagoras were
familiar with the arithmetic, geometric, and harmonic progressions. In modern
notation, Boethius definitions are as follows. The numbers a < b < c are in
arithmetic progression if b  a ¼ c  b, geometric if b/a ¼ c/b, and harmonic when
(c  b)/(b  a) ¼ c/a. Besides the property of arithmetic progressions that could be
stated, in modern notation, as n(n þ 2r) < (n þ r)2, Boethius characterized the three
progressions this way: given three numbers a < b < c in progression we have b/
a > c/b if the progression is arithmetic, b/a ¼ c/b if it is geometric, and b/a > c/b if it
is harmonic.
Given two outer numbers (the extremes), Boethius provided methods for
Downloaded by [University of Arizona] at 03:33 29 July 2012

obtaining the middle term to form a triple of numbers in progression. Thus given a
and c, we find b using the formulas:
pffiffiffiffiffi
b ¼ ða þ cÞ=2 b ¼ ac b ¼ aðc  aÞ=ða þ cÞ þ a;
for arithmetic, geometric, and harmonic progressions, respectively.
Boethius also introduced seven other progressions, as Nicomachus had done
before him. The main reason for this seems to have been to aim at the Platonic total
of 10. Here we give their definitions and examples provided by Boethius.

Name Definition Example

Fourth (b  a)/(c  b) ¼ c/a 3–5–6

Fifth (b  a)/(c  b) ¼ b/a 2–4–5
Sixth (b  a)/(c  b) ¼ c/b 1–4–6
Seventh (c  a)/(b  a) ¼ c/a 6–8–9
Eighth (c  a)/(c  b) ¼ c/a 6–7–9
Ninth (c  a)/(b  a) ¼ b/a 4–6–7
Tenth (c  a)/(c  b) ¼ b/a 3–5–8

Rithmomachia seems to have been invented in the eleventh century as a

pedagogical device to help teach Boethius’ Arithmetic. It spread through the north of
Europe, reaching Italy only in the fifteenth century. Its users were mainly the
students, not only of arithmetic, but also of astronomy and astrology, who later took
the game to other places. Some calculators paid attention to it as well. Besides its
pedagogical value, people found other qualities in the game, like a positive influence
on the spirit, or helping to fight nostalgia, which connected the game with medicine.
However, its main role was as a pedagogical game, being a kind of ‘application’ of
Boethian arithmetic.
Moyer (2002) is the best reference for understanding the route along which this
game travelled in Europe, sharing its destiny with the quadrivium itself. Bell (1960)
classifies board games into six categories: race games, war games, positional games,
Mancala games, dice games, and calculation games. These last, Bell clarifies, are
based on the Pythagorean philosophy of numbers. The only example he provides is
Rithmomachia. This philosophical attribute of the game clearly characterizes it.
 92 BSHM Bulletin

Chess, some five centuries older, had to wait still another five centuries to become the
most played game among the European elites. One description of Rithmomachia is
found in Barozzi (1572), of which the Portuguese National Library owns a copy. For
more versions of the rules, see Coughtrie (1984) and Borst (1986), which compile
several manuscripts about the game.
Here we present one of the main variants. The game unfolds on a 8  16
checkered board, where two adversaries, White (even) and Black (odd), control their
armies of 24 numbered pieces (Figure 15; see also Figure 17). The starting position
shows several relations and proportions. We show how to implement the first of the
two positions below. The second, the starting position in Rithmomachia, is easily
obtained from the first one.
Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 15. Starting position of Rithmomachia

We treat the disposition of the white pieces, the black ones follow a similar
pattern. On the right we write modern notation to explain the operations that occur.
From the even numbers 2, 4, 6, 8 placed on circular pieces in the central cells of the
sixth row, we deduce the remaining numbers. The second row, with circular pieces,
shows the squares of the numbers in the first one.

2 4 6 8 n
4 16 36 64 n2

The following row, with triangular pieces, shows numbers obtained by addition of
the previous two.

2 4 6 8 n
4 16 36 64 n2
6 20 42 72 nðn þ 1Þ
The next row also has triangular pieces. Each piece shows a number that is the
square of one more than the number on the top disc.

2 4 6 8 n
4 16 36 64 n2
6 20 42 72 nðn þ 1Þ
9 25 49 81 ðn þ 1Þ2
 Volume 26 (2011) 93

Figure 16. Rithmomachia pieces in Barozzi’s manual

Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 17. Rithmomachia board and pieces (from the exhibition Mathematical games throughout the
ages, Museum of Science, University of Lisbon)

The pieces in the first row of square shapes show numbers obtained from addition of
the previous two.
2 4 6 8 n
4 16 36 64 n2
6 20 42 72 nðn þ 1Þ
9 25 49 81 ðn þ 1Þ2
15 45 91 153 ðn þ 1Þð2n þ 1Þ
The last operation is shown below.
2 4 6 8 n
4 16 36 64 n2
6 20 42 72 nðn þ 1Þ
9 25 49 81 ðn þ 1Þ2
15 45 91 153 ðn þ 1Þð2n þ 1Þ
25 81 69 289 ð2n þ 1Þ2

There are two pieces (91 and 190), one of each colour, that are replaced by
special, solid objects, the pyramids (Figure 16). These represent sums of square
numbers, each corresponding to one of the described shapes. In this version of the
rules, the white pyramid has six storeys, showing 1, 4, 9, 16, 25, 36. The black has
five, showing 16, 25, 36, 49, 64. Note that 1 þ 4 þ 9 þ 16 þ 25 þ 36 ¼ 91 and
16 þ 25 þ 36 þ 49 þ 64 ¼ 190.
 94 BSHM Bulletin

The movements are easy: discs move orthogonally one square, triangles two cells
diagonally, while squares move three steps, orthogonally. Pyramids move as the
Queen in chess, but are limited to four squares. Captures can happen in several ways.
(i) Equality: if a piece moves to a position from where it could, in a move, reach a cell
occupied by an adversary piece with the same number, the latter is taken, without
replacement. (ii) Arithmetic operation: if two pieces belonging to the same player can
move to a cell occupied by an adversary and if the number of the latter is equal to the
sum or difference of the numbers on the pieces of the first player, the first player
captures the adversary’s piece, without replacement. (iii) Siege: if a move surrounds an
enemy piece in such a way that it cannot move, it is captured. (iv) Pyramids can be
captured by slices or whole. There are a few extra rules regarding the capture of
pyramids that we will skip here. Captured pieces that are replaced are turned and
reintroduced into play immediately. A player wins when, in the adversary’s half
Downloaded by [University of Arizona] at 03:33 29 July 2012

board, he places three pieces into progression (any of the three, for instance, 2–15–18,
9–15–25, 9–15–45), or four pieces showing a combination of two progressions
(2–3–4–8, 3–4–5–6, 2–3–6–12), or even all three (4–6–9–12), progressions.
Following closely the fate of the quadrivium, Rithmomachia vanished, leaving no
trace. Chess took over as the most popular game in Europe. Its history shows how
culturally relevant a pedagogical game can be, and how much we can learn from a
game about the science and teaching of the past.
In his De viribus quantitatis (c. 1500) (see Singmaster 2008), the oldest book on
recreational mathematics, Luca Pacioli described the Chinese Rings (Figure 18), a
topological puzzle still popular today.

Figure 18. A modern version of the Chinese Rings puzzle

Peg Solitaire (Figure 19) traces its origins from seventeenth-century France. It is a
game where a board has all its holes occupied with pegs except for the central one.
The goal is, making valid moves (small jump capture), to empty the entire board but
for a solitary peg in the central hole.
To stress the respect with which some games have been regarded, we quote from
Thomas Moore’s description of an ideal society (Moore 2010):
They know nothing about gambling with dice or other such foolish and ruinous
games. They play two games not unlike our chess. One is a battle of numbers, in
which one number captures another. The other is a game in which the vice battles
against the virtues.
 Volume 26 (2011) 95

Figure 19. Peg Solitaire: starting and target position

Downloaded by [University of Arizona] at 03:33 29 July 2012

We are convinced that Moore was referring to Rithmomachia. The other game, with
vice and virtues, is still a mystery. It is not Ludus Regularis (in which no vice occurs).
It might be Preceitos Morais (Figure 20), a game invented by the Portuguese
historian João de Barros, who published a book about it in 1540 (Barros 2009), but
we are not certain. Utopia appeared in 1516, but it is not impossible that Moore had
learned about the game well before Barros published his book.

Figure 20. Board and spinner of Preceitos Morais (author’s collection)

Ludus Astronomorum (Figure 21), a game for seven players based on Ptolemaic
astrology was addressed in Alfonso’s Book of Games as well. This game, introduced
to Europe by the Arabs, was a representation of the sky and its seven ‘planets’. The
seven concentric rings (for the Moon, Mercury, Venus, Sun, Mars, Jupiter, and
Saturn) are inscribed in a Zodiac with twelve houses. Depending on the house in
which a piece lands (according to the throw of a die) a player receives 24 from each
player 2 houses away, 36 from each 3 houses away, but pays 36 to each player 4
houses away, and so on. The game encourages players to recognize the astrological
configurations.
 96 BSHM Bulletin
Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 21. Ludus Astronomurum in Alfonso’s Book of Games

Ludus Astronomorum was popular for several centuries. Based on it, William
Fulke, in the sixteenth century, invented another astronomical game. Fulke (1571)
named it Ouranomachia (Figure 22), and meant it to be a scientifically updated
version of the astronomy game.

Figure 22. Ouranomachia board (from the exhibition Mathematical games throughout the ages, Museum
of Science, University of Lisbon)

Besides writing a Rithmomachia manual (1563) Fulke invented another game

(1578), inspired by the numerical game, intended to be a pedagogical tool to teach
geometry, Metromachia (Figure 23). In this extremely complicated game the
behaviour of the pieces was determined by their geometrical properties (shape,
perimeter, area, volume).
 Volume 26 (2011) 97
Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 23. A page of Metromachia

In 1697, Joseph Moxon published a book in London that he claimed René

Descartes (1697) had written, and which he had translated himself, Use of pedagogical
playing cards (Figure 24). All the illustrations of the proofs were put on 52 playing
cards at the end of the book. The King of Diamonds, for example, illustrates the
constructions of perpendiculars to a segment (Figure 25). This is a work in geometry,
with a Euclidean flavour but different axioms. Here we can find these
1 – It is possible to trace a segment from point A to point B.
2 – It is possible to extend a segment.
3 – It is possible to trace a circle centred at A with radius AB.
4 – Given two points it is possible to construct a section.
In De Ludo Algebraico, George Berkeley, the Irish philosopher, described a game
he was proud to have created to teach Algebra (Figure 26) (see Berkely 2007). To
support his trust in the pedagogical virtues of games and algebra, Berkeley quoted
some respected authors, like Francis Bacon (The advancement of learning):
So that as tennis is a game of no use in itself, but of great use in respect it maketh
a quick eye and a body ready to put itself into all postures; so in the Mathematics,
that use which is collateral and intervenient is no less worthy than that which is
principal and intended.
 98 BSHM Bulletin
Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 24. Use of the geometrical playing cards

Figure 25. King and 10 of Hearts, King of Diamonds

and John Locke (On the conduct of understanding):

I have before mentioned mathematics, wherein algebra gives new helps and views
to the understanding. If I propose these, it is not, as I said, to make every man a
thorough mathematician or a deep algebraist; but yet I think the study of them is
of infinite use even to grown men.
Berkeley agreed with using games in teaching (‘and has gained every point, who
has combined the useful with the agreeable’) and believed algebra was the best
subject to teach (‘It is styled by all, the great, the wonderful art, the highest
pinnacle of human knowledge, the kernel and key of all mathematics; and, by
some, the foundation of all sciences’). Ludus Algebraicus uses a board and a
 Volume 26 (2011) 99
Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 26. Ludus Algebraicus board

moving piece to generate algebraic equations. Two students then compete to solve
the equations and systems.
Charles Dodgson invented a game called Logic (Figure 27) and published a book
on it (Carroll 2006) under the name of Lewis Carroll. The purpose of the game was
to draw conclusions, in a syllogistic way, by means of moving counters on a board.
His book contains many examples and exercises.

Figure 27. The game of Logic (author’s collection)

William Hamilton, an Irish mathematician, invented the Icosian Game (see also
the article by Genovese et al. in this issue). In one version of this puzzle the task
consists of finding a path that visits each vertex exactly once. The concept of a
Hamiltonian graph arose from this game. From the exhibition Mathematical games
 100 BSHM Bulletin

throughout the ages, Figures 28 and 29 show two replicas of Hamilton’s own
implementations:
Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 28. The Icosian game as a ‘rattle’ puzzle

Figure 29. The Icosian game as a board game

The Towers of Hanoi (Figure 30) is a puzzle invented in 1883 by M Claus, an

anagram of Edouard Lucas (see Lucas 1979). A pile of discs of decreasing radius is
stacked on one of three poles. Moving one disc at a time, without letting a bigger disc
rest on a smaller one, we are asked to change the pile from one pole to another. The
recursive character of the solution to this puzzle makes it somewhat similar to the
Chinese Rings.
Tangram spurred the first world puzzle craze. The nineteenth century witnessed
another one with a puzzle named ‘15’ (Figure 31) (Slocum and Sonneveld 2006). Sam
Loyd, the great American puzzle inventor, claimed to be its creator, which turned out
to be false. It consists of a sliding device, a 4  4 array with the numbers 1 to 15, and
an empty cell. The aim is to bring back a scrambled position back to the natural order.
 Volume 26 (2011) 101

Figure 30. Towers of Hanoi: starting and target positions

Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 31. The puzzle 15 target position

Sam Loyd offered USD$1,000 to whoever could solve the position shown in
Figure 32. The prize was never claimed! The impossibility of this challenge is better
understood when phrased in the language of group theory.

Figure 32. An impossible task. . .

Hex (Figure 33) was invented twice in a short period of time, by Piet Hein, a Danish
scientist and poet, in 1942, and by the mathematician and Nobel laureate John Nash in
1947. The rules of Hex are surprisingly simple: players alternate in placing counters
of different colours, to try to connect parallel margins of the board. That there
can be no ties in this game can be seen in several ways, of varying formal presentation.
 102 BSHM Bulletin

Figure 33. Empty Hex board

Downloaded by [University of Arizona] at 03:33 29 July 2012

If we assume the board filled with red and blue pieces, and think of blue as water and
red as one-foot tall walls and bricks, then either water flows between blue margins
(blue wins) or there is a red deck preventing blue communication (red wins). Proofs
with graph theory and by induction can be found in Cameron Browne’s book (2000).
This result is known as the Hex Theorem.
Nash gave a beautiful non-constructive proof that the first player must have a
winning strategy in Hex. His argument, the now famous ‘strategy stealing argument’,
goes as follows. As there are no ties, either first or second player must have a winning
strategy. Suppose it is the second. Then the first player makes his first move at
random and assumes the role of second player. The reason this trick works is that,
Hex being a connection game, an extra move, no matter how bad it might be, cannot
hurt its player. With this plan, therefore, the first player wins the game, which gives
us a contradiction. We must conclude that the first player has a winning strategy. If
the dimension of the board is not too small, nobody knows how to play perfectly, so
the practice of Hex is not ruined by Nash’s result.
The connection between Hex and mathematics was emphasized when David
Gale, a mathematician from Berkeley, showed that Hex Theorem is logically
equivalent to Brouwer’s Fixed Point Theorem, a deep result in topology (Gale 1979).
Knowing this equivalence does not increase one’s play strength, but shows how
abstract games and mathematics are linked in a fascinating way.
Another mathematical puzzle that infected the world was the Rubik Cube
(Figure 34), created by the Hungarian architect Erno Rubik in the 1970s, which
became the best selling puzzle in history. A 3  3  3 cube, with differently coloured
faces, moves by slices, getting scrambled with just a few moves. To find the way back
to the starting position is an incredible challenge. This toy puzzle is used to illustrate
many group theory concepts. Conversely, knowledge of group theory facilitates
understanding of the puzzle itself.
The work of Leonhard Euler on Latin Squares (arrays of symbols with no
repetitions in rows or columns) is the remote conceptual background for the now
ubiquitous Sudoku puzzle (Figure 35), which appeared in an American magazine
in the 1970s but became famous first in Japan and then throughout the world.
Played with numbers, Sudoku is a logical challenge, not an arithmetical one.
Any set of nine symbols would do, numbers being just universally familiar. The
task consists on filling the blanks with the digits 1 to 9 without row, column, or
box repetitions.
 Volume 26 (2011) 103
Downloaded by [University of Arizona] at 03:33 29 July 2012

Figure 34. A Rubik cube unscrambled

Figure 35. Sudoku puzzle

I hope I have managed to show that the relation between mathematics and games
is extremely varied and hard to tame. These two realms keep their secrets. It is our
task to try to uncover them. In this metagame we are fortunate to enjoy the pleasure
of abstract thinking at its highest level!

Bibliography

Alfonso X, Los libros de acedrex dados e tablas, Fundacion Jose Antonio de Castro, 2007.
Barozzi, F, Il nobilissimo et antiquissimo Givoco Pythagoreo nominato Rythmomachia cioe
Battaglia de Consonantie de Numeri, Venice, 1572.
Barros, J, Diálogo de Preceitos Morais em Forma de Jogo, Apenas, 2009.
Bell, R C, Board and table games from many civilizations, Oxford University Press, 1960.
Berkeley, G, ‘De Ludo Algebraico’, The Works of George Berkeley, Vol. IV, Elibron, 2007.
Berlekamp, E, Conway, J and Guy, R, Winning ways for your mathematical plays, Vol. I–IV,
AK Peters, 2001.
Berlekamp, E and Wolfe, D, Mathematical Go: chilling gets the last point, AK Peters, 1994.
Boèce, Institution Arithme´tique, Les Belles Lettres, 1995.
Borst, A, Das mittelalterliche Zahlenkampfspiel, Carl Winter Universitaetsverlag, 1986.
Browne, C, Hex strategy: making the right connections, AK Peters, 2000.
Carroll, L, The game of Logic, Hard Press, 2006.
 104 BSHM Bulletin

Cornford, F, Plato’s cosmology: ‘Timaeus’ of Plato, Hackett Publishing, 1997.

Coughtrie, M E, Rhythmomachia, a propaedeutic game of the middle ages, University of Cape
Town, 1984.
De Voogt, A, Mancala board games, British Museum Press, 1997.
Descartes, R, The use of the geometric playing cards, London, 1697.
Euler, L, ‘Solution d’une question curieuse que ne paroit soumise à aucune analyse’, Me´moires
de l’Acade´mie des sciences de Berlin, 15 (1759) [1766], 310–337.
Fulke, W, The most noble aucient, and learned playe, called the Philosophers game, invented for
the honest recreation of students, and other sober persons, in passing the tediousnes of
tyme, to the release of their labours, and the exercise of their wittes, London, 1563.
Fulke, W, Ouranomachia hoc est astrologorum ludus, London, 1571.
Fulke, William, Metromaxia sive ludus geometricus, London, 1578.
Gale, D, ‘The game of Hex and the Brouwer Fixed-Point Theorem’, American Mathematical
Monthly, 86 (1979), 818–827.
Le, Glay, Chronique d’Arras et de Cambrai par Balderic, Paris, 1834.
Downloaded by [University of Arizona] at 03:33 29 July 2012

Hardy, G, A mathematician’s apology, Cambridge University Press, 1992.

Lhôte, J -M, Histoire des jeux de socie´te´, Flammarion, 1993.
Lucas, E, Re´cre´ations mathe´matiques, Vol. 3 (1892), Librairie Albert Blanchard, 1979.
Moore and Thomas, Utopia, Norton, 2010.
Moyer, A E, The philosopher’s game: Rithmomachia in Medieval and Renaissance Europe, The
University of Michigan Press, 2002.
Murray, H J R, A history of chess, Oxford University Press, 1913.
Murray, H J R, A history of board games other than chess, Oxford University Press, 1951.
Netz, R and Noel, W, The Archimedes Codex: revealing the secrets of the world’s greatest
palimpsest, Weidenfeld and Nicolson, 2007.
Nicomachus, Introduction to arithmetic, Britannica Great Books, 11, 1952.
Pegg, E, The Loculus of Archimedes, solved, http://www.maa.org/editorial/mathgames/
mathgames_11_17_03.html
Petkovich, M, Mathematics and chess, Dover, 1997.
Schaeffer, J, et al., ‘Checkers is solved’, Nature, (2007).
Singmaster, D, ‘De Viribus Quantitatis by Luca Pacioli’, in A lifetime of puzzles, E Demaine,
M Demaine, and T Rogers (eds), AK Peters, 2008.
Slocum, J, The Tangram book: the story of the Chinese puzzle with over 2000 puzzles to solve,
Sterling Publications, 2003.
Slocum, J and Sonneveld, D, The 15 puzzle: how it drove the world crazy; the puzzle that started
the craze of 1880; how America’s greatest puzzle designer, Sam Loyd, fooled everyone for
115 years, Slocum Puzzle Foundation, 2006.
Watkins, J, Across the board: the mathematics of chessboard problems, Princeton University
Press, 2007.