THE SOCCER WORLD CUP

A Mathematical Approach

By: Andrei C. Grecu

April 17, 2006
ORFE Advisor: John Mulvey

Submitted in partial fulfillment

Of the requirements for the degree of
Bachelor of Science in Engineering
Department of Operational Research and Financial Engineering
Princeton University

1
I hereby declare that I am the sole author of this thesis.

I authorize Princeton University to lend this thesis to other institutions or individuals for
the purpose of scholarly research.

Andrei C. Grecu

I further authorize Princeton University to reproduce this thesis by photocopying or by

other means, in total or in part, at the request of other institutions or individuals for the
purpose of scholarly research.

Andrei C. Grecu

2
To my grandfather, Nicolae Dumitrescu

3
 ACKNOWLEDGEMENTS

I am especially grateful to Professor Mulvey for his guidance in writing this

thesis, his advice and his unyielding support. I would also want to thank Professor

Vanderbei, Professor Powell, and Professor Lord for teaching me a lot of the material

used in this thesis. In addition, I am indebt to the whole Princeton community for

challenging my mind over the last four years.

Dear family, thank you for your incredible support from thousands of miles away,

in Romania. I would have been lost without you in this process.

4
 TABLE OF CONTENTS

CHAPTER 1: INTRODUCTION……………………… 6 – 11

CHAPTER 2: RANKING ALGORITHMS………… 12 – 41

2.1 EIGENVECTOR RANKINGS……………. (16-24)

2.2 RANDOM WALKERS RANKINGS……… (25-30)

2.3 NEURAL NETWORK RANKINGS……… (31-41)

CHAPTER 3: A POISSON MODEL FOR

WORLD CUP GAMES………….………………42 – 57

CHAPTER 4: SIMULATING THE WORLD CUP…58 – 65

CHAPTER 5: CONCLUSION…………………………66 – 69

APPENDIX A: NEURAL NETWORK C++ CODE…70 –75

APPENDIX B: NEURAL NETWORK EVOLUTION 76

BIBLIOGRAPHY……………………………………… 77 – 78

5
Chapter 1: Introduction Andrei C. Grecu

Chapter1: Introduction

“ Some people believe soccer is a matter of life and death.

I am very disappointed with that attitude.

I can assure you it is much, much more than that. “

Bill Shankly, President of F.C. Liverpool

The Soccer World Cup tournament, held every four years, is the most watched

sports event in the world, surpassing even the Olympic Games. Moreover, enthusiasm for

soccer is relatively uniformly spread across the world, both in term of spectators and

actual participants. For example, 194 national teams, essentially a team for every country

in the world, competed in qualifying games for a place in the next World Cup final

tournament, which will take place in Germany, in June-July 2006. Given the international

appeal of soccer and of the World Cup in particular, there is a huge interest in assessing

the relative strengths of the teams that take part in the tournament and implicitly

predicting the overall winner.

The intrinsic complexity of the World Cup tournament makes the job of ranking

participating teams both challenging and mathematically stirring. In the first phase of the

tournament, over a period of about two years, approximately 200 national teams compete

in qualifying games on five continents, with the highest ranked teams in each

6
Chapter 1: Introduction Andrei C. Grecu

geographical region advancing to the final tournament. Subsequently, the 32 qualified

teams travel to the country that organizes the World Cup and embark upon a marathon of

games. Initially, the 32 teams are divided into eight groups of four, with each group

playing six round robin matches. Following the initial games, the top two teams from

each group advance to the second round of the final tournament, which consists of a

simple 16-team knockout tournament. Thus, after approximately one month and 64

matches played, the final of the World Cup determines the World Cup Champion, “the

best” national team in the world for the next four years.

While the World Cup effectively compares teams by having them play against

one another on the soccer field, any attempt to rank the teams before the start of the

competition faces a number of challenges. First, by its very nature soccer is a highly

unpredictable game- only a few goals are scored during the ninety minutes of a normal

game, and as opposed to baseball or football, few statistics are recorded for each game.

Second, participating teams come from various regions of the world, in which soccer is

played at different levels- even though the winner of a geographic region might be much

weaker than the winner of another region, the format of the World Cup guarantees that

teams from all regions take part in the final tournament. Third, the relative rarity of

games between national teams makes it difficult to find sources for comparison- a soccer

national team only plays an average of ten games every year, so there are still national

teams that have never played one another. Despite the challenges discussed here, a

number of papers have tried to model soccer scores and simulate the World Cup

tournament.

7
Chapter 1: Introduction Andrei C. Grecu

Literature Review

Kuonen (1997) fits a logistic regression for the probability of winning a soccer

game, based on seeding coefficients computed from the outcome of previous games

played by the two teams. Karlis and Ntzoufras (1998) examine the choice of a Poisson

distribution for modeling goals in soccer games and build a Poisson log-linear model for

scores in the Greek national league. Dyte and Clarke (2000) also assume a Poisson

distribution of goals and model the 2002 World Cup based on the (controversial) national

team rankings provided by the Federation of International Football Association (FIFA).

The Palomino, Rigotti, and Rustichini (1998) model, inspired from game theory, looks at

a soccer game in continuous time and examines the effect of three factors- team’s ability

(performance record), passion (home-field advantage), and strategy (reaction to current

score), in determining the outcome of a soccer game. Finally, Koning (2001) simulates

soccer championships using a Poisson fit to predict game outcomes based on historical

scoring intensities of the two teams involved in the game.

Simulating the soccer World Cup final tournament is even more challenging

than simulating the games in a national league. Teams that compete in the World Cup

final tournament come from different qualifying tournaments, in which the soccer played

can have different characteristics. For example, national teams in South America tend to

score on average more goals when they play each other than teams in Europe do.

Therefore, the average number of goals scored per game might not be the best variable to

predict the result of an encounter between a team in South America and a team in Europe.

In all the models mentioned in the previous paragraph, the variables used in fitting the

regressions are derived from aggregate statistics of the previous performance of the two

8
Chapter 1: Introduction Andrei C. Grecu

teams. This approach might be reasonable for teams competing against the same

opponents in a league, but in the case of the World Cup we need to adjust the variables to

account for the different backgrounds from which participating teams come from.

Regarding the soccer World Cup as a competition between teams coming from

different regional tournaments brings to mind the problem faced every year by the Bowl

Championship Series (BCS) in ranking American college football teams. The two

problems- ranking college football teams and ranking World Cup soccer teams, are

similar in that: i) the number of games played by every team is relatively small, ii) teams

play many more games within their regional league than against teams in other leagues,

and iii) crucially, the quality of the opponents of different teams varies from region to

region. Thus, the techniques used to rank college football teams can be tailored to rank

the teams that participate in the soccer World Cup.

The theory behind ranking teams in uneven paired competitions has strong and

diverse mathematical foundations. Keener (1993) formulates the ranking problem as a

linear eigenvalue problem and solves it by using the result of the Perron-Frobenius

theorem. Goddard (1983), Stob (1985) and Ali, Cook, and Kress (1986) develop

algorithms for rankings that satisfy the so-called minimum violations ranking (MVR)

criterion, which minimizes the instances in which lower ranked teams defeat higher

ranked teams. Wilson (1995) builds a neural network based on previous interactions

between teams and looks for the equilibrium values of the network. Finally, Thomson

(1975) and Reid (2003) design least squares and maximum likelihood methods for

ranking teams. Even though some these rankings have strong mathematical foundations,

9
Chapter 1: Introduction Andrei C. Grecu

all ranking methods are intrinsically subjective by means of the variables chosen to

explain outcomes, the parameters of the model, and the final interpretation of the results.

Overview

This thesis looks at the soccer World Cup and uses mathematical tools to

determine the relative strengths of the participating teams and simulate the structure of

games within the tournament. Given the various backgrounds of the qualified teams, and

the complexity of the World Cup tournament, I use a three-step approach to assessing the

relative strengths of the national soccer teams that participate in the soccer World Cup.

First, in Chapter 2 I implement three algorithms to rank the participating teams,

before the start of the tournament: i) a matrix-based method takes as input previous

interactions among teams and returns an eigenvector with the relative value assigned to

each team. ii) a random walker algorithm looks at the steady-state macroscopic solution

of a setting in which a number of vacillating voters perpetually change their mind

regarding their favorite team, thus executing random walks on a network defined by the

participating teams (nodes) and their previous interactions (edges). iii) a neural network

algorithm looks at a neural network whose nodes are the participating teams and whose

connections are determined by previous interactions among the teams, and uses a soccer-

intuitive transfer function to update the value of each node (team) until a steady state

solution is reached.

Second, in Chapter 3 I use the rankings presented above to make predictions

regarding the outcome of separate games in the World Cup. I start by assuming a Poisson

distribution of goals scored in a soccer game and I fit a nonlinear regression using the

10
Chapter 1: Introduction Andrei C. Grecu

number of goals scored by each team in the games from the 2002 World Cup. Using the

regression results, I compute the probability of winning assigned to both teams

competing in a World Cup game. However, given the complex structure of the soccer

World Cup, it is difficult to explicitly calculate the conditional probabilities for each team

to win the tournament, so I decide to simulate the games instead.

Third, the simulation of the World Cup games performed in Chapter 4 allows me

to determine the probability of winning the World Cup assigned to each team. In order to

determine the manner in which teams are favored or disadvantaged by the World Cup

draw, I also simulate a round robin competition between all the teams qualified for the

World Cup. Even though such full competition is not possible in practice because of the

great number of games involved, it is considered intuitively the fairest way of

determining the best team. Therefore, by comparing the winning probabilities from the

World Cup simulation with the winning probabilities from the round robin simulation, I

am able to determine which teams had a “lucky draw” for the World Cup games.

Finally, I conclude in Chapter 5 by examining the accuracy of my predictions for

the 2002 World Cup, making predictions for the up-coming 2006 World Cup, and

drawing an analogy between sports betting and the financial markets.

11
Chapter 2: Ranking Algorithms Andrei C. Grecu

Chapter 2: Ranking Algorithms

Overview

This chapter presents three algorithms used for ranking the 32 soccer national

teams that qualified for the 2006 Soccer World Cup. By looking at historical games

played between soccer national teams around the world, each algorithm returns the

relative values of the teams prior to the start of the tournament. Whilst past performance

is not always a good indicator of the present value of a team, I will focus on those

historical games that are most relevant to the real value of the teams.

First, I look at games that took place in conditions somehow similar to the World

Cup, namely games played on neutral field as part of a relevant, competitive

championship. Luckily, independent of the World Cup, national soccer teams also

compete for regional supremacy on each continent. Thus, every two years, African

countries compete in the African Nations Cup. Every four years, national teams in Asia

compete for the Asian Cup. Held every two years, the CONCACAF Gold Cup reunites

teams in North America. Every three years, teams in South America take part in Copa

America. Finally, every four years the best teams in Europe participate in the European

Championship. This plethora of regional tournaments generates a significant number of

interactions between teams on the same continent. Conveniently, the World Cup, which

takes place every four years, reunites teams from all over the world, thus allowing

comparisons between teams on different continents.

12
Chapter 2: Ranking Algorithms Andrei C. Grecu

Second, I am interested in historical games played between two teams of

comparable strength. For ranking purposes, the fact that Australia beat Cook Islands 16-0

is far less relevant than a tight game between two teams of comparable strengths, such as

Germany and Argentina. Therefore, I will focus my analysis only on games in which both

teams qualified for at least one of the previous three World Cups. This totals 55

competitive teams, with 26 teams from Europe, 9 from Africa, 8 from South America, 7

from Asia, and 5 from North America:

North
America Europe
(5) (29)

Asia
(7)

South
America Africa
(8) (8)

Figure 1: The five qualifying regional tournaments with the number of

teams that qualified to at least one World Cup since 1994 in parentheses

Even though only 32 soccer teams qualified for the 2006 World Cup, I choose to

look at games played among 55 teams of comparable strength. By looking at more games

and ranking 55 instead of only 32 teams, I increase the precision of the algorithms in

determining the values of the 32 teams that did qualify for the World Cup. For example,

even though Cameroon did not qualify for the World Cup this year, its previous games

13
Chapter 2: Ranking Algorithms Andrei C. Grecu

against competitive teams around the world allows us to better rank other African teams.

For instance, Togo qualified for the first time to the World Cup this year, so it does not

have direct previous interactions with teams on other continents. However, its games

against Cameroon and other African countries allow us to better determine Togo’s overall

strength.

Finally, I consider games played between soccer national teams over the last 12

years. Although this is a long period of time, each game is also assigned a weight that

decreases with the number of years since the game took place. Thus, a game that took

place ten years ago is approximately ten times less significant in determining the current

value of the team than a game that took place this year is. In addition, a lot of skilled

soccer players start playing for their national teams when they are very young and they

do play for ten or twelve years before they retire. Also, it is not unusual for a national

team to form around a nucleus of talented players who will play for their country for

approximately a decade. Consequently, putting heavy weights on recent games, but also

decreasing weights on past encounters, effectively captures the relative strengths of the

teams over time. By comparison, the Federation of International Football Association

(FIFA) also uses results from the previous eight years in computing its coefficients for

each country.

To sum up, I look at historical games between competitive teams as part of a

competitive tournament, in which the teams play their best players at full potential. In

particular, I look at 477 games among 55 teams at the following competitions:

14
Chapter 2: Ranking Algorithms Andrei C. Grecu

1. World Cup: South Korea/Japan 2002, France 1998, USA 1994

2. African Nations Cup: Egypt 2006, Tunisia 2004, Mali 2002, Ghana 2000,

Burkina Faso 1998, South Africa 1996, Tunisia 1994

3. Asian Cup: China 2004, Lebanon 2000, United Arab Emirates 1996

4. CONCACAF Gold Cup: USA 2005, Mexico 2003, USA 2002, USA 2000,

USA 1998, USA 1996

5. Copa America: Peru 2004, Colombia 2001, Paraguay 1999, Bolivia 1997,

Uruguay 1995

6. European Championship: Portugal 2004, Belgium/Netherlands 2000, England

1996

Acknowledgements

The first algorithm is inspired from the matrix-based algorithms for ranking

American college football teams, as discussed in Keener (1993), Boginski, Butenko, and

Pardalos (2004), and Martinich (2003). The second algorithm is a variation of the random

walker algorithm discussed in Callaghan, Mucha, and Porter (2005). The third algorithm

is a neural network method somehow similar to the approach presented in Wilson (1995).

However, I modify each of these algorithms in at least two important ways. First,

I implement soccer-intuitive functions and parameters in each algorithm, so that they deal

with results of soccer rather than football games. Second, since I look at games that took

place over a longer period of time, I discount the importance assigned to each game, so

that recent games have a much greater impact on the rankings than older games do.

15
Chapter 2: Ranking Algorithms Andrei C. Grecu

2.1 Eigenvector Rankings

The Intuition

“Spain beat Italy, and Italy beat Brazil, therefore Spain should win against

Brazil” or more intricate “Brazil tied against Germany, Germany beat Japan, and Brazil

lost against Italy, therefore Italy should definitely win against Japan”- sport fans often

make such kinds of conjectures regarding the outcome of future games based on previous

results. Even though such predictions often turn out to be wrong, the deduction process is

not entirely speculative, as previous games do contain information regarding the relative

strengths of the teams. For example, consider the following history of games between

soccer national teams represented by a directed graph- an arrow from BRA to GER

indicates that team BRA beat team GER:

BRA GER

TUN
ITA
SPA

JAP

Figure 2: Directed graph showing a hypothetical history of games

In this scenario, JAP clearly looks like the weakest of the six teams depicted, but

it is less intuitive how the other teams are ranked, especially since different teams have

16
Chapter 2: Ranking Algorithms Andrei C. Grecu

played a different number of games. A starting point for comparison might be the

number of games won

winning percentage of each team, winning% = :
total number of games

BRA GER ITA SPA TUN JAP

0.500 0.667 0.500 0.667 0.500 0.000

Using winning percentage as the ranking criterion, GER and SPA are tied for first

place. However, such a ranking does not take into account the quality of teams defeated

by each team Should GER and SPA get the same credit for defeating ITA and TUN,

respectively? In order to account for the strength of schedule of each team, we calculate

updated winning percentages using the defeated team’s winning percentage instead of a

1, to account for a victory,

 (winning% of defeated teams) . For example, the updated
total number of games

0.500  0.667
winning percentage for GER is  0.389 , since GER played a total of three
3

games and beat ITA and SPA, whose winning percentages were 0.50, and 0.67,

respectively. Similarly, the updated winning percentage for SPA, which used to be tied

(0.500  0.500)
with GER, is now  0.333 . With the updated winning percentages, GER
3

is ranked ahead of everybody else since it beat better opponents:

BRA GER ITA SPA TUN JAP

0.333 0.389 0.125 0.333 0.000 0.000

17
Chapter 2: Ranking Algorithms Andrei C. Grecu

However, if we update the winning percentages once again, we obtain an even

better ranking, based upon already updated winning percentages for each team. With the

new percentages, BRA is now ranked in first place! However, if we keep updating the

winning percentages following the process described above, the rankings eventually

stabilize to the following:

BRA GER ITA SPA TUN JAP

0.344 0.290 0.204 0.162 0.000 0.000

In other words, BRA gets the first place because its victory came against a very

strong GER, while GER’s two victories came against relatively weaker ITA and SPA.

Also, even though ITA and SPA have the same number of victories as GER, half their

victories came against the weak teams TUN and JAP, which in turn lowered their own

ranking. Thus, using a reasonably intuitive algorithm, we were able to rank the teams

according to their winning percentages adjusted for the relative strength of their

schedules.

The Mathematical Model

Consider a competition in which N participants play an uneven paired schedule,

meaning that not all teams play each other. Let ni be the number of games played by

participant i, and let aij be some nonnegative number assigned to each team, depending on

the outcome of the game between participant i and participant j. Now, if we assume that r

18
Chapter 2: Ranking Algorithms Andrei C. Grecu

is a vector of ranking values, with positive component rj indicating the strength of the jth

participant, then we define the overall score for participant i as:

1 N
si  *  aij * rj
ni j 1

The division by ni prevents teams from accumulating a large score just by playing

extra games. Also, the matrix A with entries aij is often called a preference or dominance

matrix, since it contains the scores assigned to each team based on its previous

interactions with other teams. If we further assume that the rank of a team is proportional

to its score, then the ranking vector r has to be a positive eigenvector of the positive

matrix A:

A* r   * r

An example of a scheme that assigns scores aij as a function of the outcome of a

separate game between team i and team j, is the following:

 1, if team i beat team j


aij =  0.5, if team i and team j tied
 0, if team i lost against team j

With the choice of aij described above and letting r0 be the column vector with j

entries, it is easy to check that A*r0 is the column vector corresponding to the winning

percentages of each team, A2*r0 gives the average winning percentage of all defeated

teams, and so on. The solution that we are looking for is the limit, lim (An*r0). However,
n

the product An*r0 gets very small as n tends to infinity, so we use the power method to

19
Chapter 2: Ranking Algorithms Andrei C. Grecu

find the ranking vector r, which is the eigenvector corresponding to the largest

eigenvalue of A:

An * r0
r  lim n
n  | A * r0 |

The Perron-Frobenius theorem tells us when this limit exists and gives a unique,

positive solution for the ranking vector r:

Theorem: If matrix A has nonnegative entries, then there exists an eigenvector r

with nonnegative entries, corresponding to a positive eigenvalue λ. Furthermore, if the

matrix A is irreducible, the eigenvector r has strictly positive entries, is unique and

simple, and the corresponding eigenvalue if the largest eigenvalue of A in absolute value.

For paired competitions, the matrix A is irreducible if there is no partition of the

teams into two sets S and T such that no team in S has played any team in T, or every

game between a team from S and a team from T resulted in a victory for the team in S. In

other words, we need all teams to be connected by previous games, and we need a cycle

in the directed graph- we need a sequence of distinct teams t1, t2, t3, …, tk such that t1 beat

t2, t2 beat t3, …, and tk beat t1. In particular, for the preference matrix to be irreducible,

there can be no winless teams. When these conditions are satisfied, the limit

An * r 0
r  lim converges to the unique positive eigenvector of A, which gives a
n  | An * r 0 |

positive ranking of teams.

20
Chapter 2: Ranking Algorithms Andrei C. Grecu

Using the preference matrix A to find the ranking vector of teams has sound

mathematical foundations and effectively takes into account the strength of schedule of

each team when rankings the teams. However, the weakness of this method comes from

the subjective choice of the score aij for team i, following a game between team i and

team j. A scheme that assigns a score of 1 for a victory, 0.5 for a tie, and 0 for a loss, is

the natural choice, since in this case A*r0 gives the winning percentages of the teams, and

A2*r0 gives the average winning percentage of all defeated teams, thus containing

information regarding the strength of schedule. In particular, this scheme works well

when teams play each other frequently, thus making aij a better indicator of the

comparative strength of the two teams. However, when interactions between teams are

relatively rare or even reduced to one game- as is often the case with soccer national

teams, this simple scheme ineffectually gives all of the credit for the win to the winner,

while the loser gets a score of zero, regardless of the degree of its defeat.

In order to avoid lopsided splits of the merit for a victory between the two teams,

I implement a formula that assigns both teams a score of 0.5 before the start of the game,

and then updates the scores, as a function of the number of goals scored by each team:

Ycf  Gij
aij 
2 * Ycf  Gij  G ji

Where:

 Gij is the number of goals scored by team i against team j



 Y is a year coefficient equal to number of year since the game took place
 cf

21
Chapter 2: Ranking Algorithms Andrei C. Grecu

In case of a tie, Gij= Gji, this scheme assigns each team a score of 0.5. Then, the

more convincing the victory, the closer the winner’s score gets to 1 and the loser’s score

gets to 0, without actually equaling 1 or 0, respectively. Furthermore, by means of the

year coefficient Ycf, this scheme puts a lot of emphasis on games played in recent years

and decreasing weights on games played a long time ago. For example, a 3-1 victory of

1 3
team i over team j brings team i aij=  0.667 points for a game played in 2006,
2  3 1

12  3
but only aij=  0.536 points for a games back in 1994. By assigning the
2*12  3  1

winning team two thirds of the one point per game in 2006, but only slightly more than

the initial half-half split in 1994, this scheme clearly rewards current success more than

past results.

Application to the 2006 Soccer World Cup

In order to rank the teams that qualified to the 2006 soccer World Cup, I applied

the algorithm described above to 477 games among the 55 national teams that took part

in at least one World Cup between 1994 and 2006. Importantly, all 477 games took place

during a major tournament, under conditions somehow similar to the World Cup. First,

teams had a great incentive to win the game and advance to the next phase of the

tournament, thus playing its best available players at full capacity. Second, games took

place on neutral territory, so that no teams (other than the nation organizing the

competition) enjoyed the home-field advantage.

Plugging the 477 games in Excel, and manipulating the data to assign a score to

all previous interactions using the scheme discussed in the previous section, I formed the

22
Chapter 2: Ranking Algorithms Andrei C. Grecu

An * r 0
dominance matrix A. Using MATLAB to find the eigenvector r  lim n ,I
n  | A * r 0 |

obtained the ranking of the 55 teams considered. Within the ranking of the 55 teams, I

identify the teams that qualified for the 2006 World Cup in Germany, and I scale the

results on a 100-point scale for the following final rankings1:

1
Australia and Ukraine did not participate in any of the 27 major tournaments considered, so they were not
among the 55 teams ranked. Consequently, I arbitrarily assumed that Australia and Ukraine take the value
of the last competitive team that they defeated: Australia defeated Uruguay in the playoffs for the 2006
World Cup, and Ukraine defeated Turkey in the 2nd European qualifying group. Therefore, I assigned
Australia the value of Uruguay, and Ukraine the value of Turkey.

23
Chapter 2: Ranking Algorithms Andrei C. Grecu

2006 EIGENVECTOR RANKING

1 BRAZIL 100.00
2 ITALY 87.52
3 GERMANY 84.65
4 MEXICO 81.89
5 NETHERLAND 77.87
6 ARGENTINA 73.33
7 FRANCE 72.42
8 ENGLAND 68.29
9 SPAIN 65.46
10 PARAGUAY 64.55
11 SWEDEN 60.96
12 CROATIA 60.78
13 PORTUGAL 58.53
14 USA 57.15
15 SOUTH KOREA 52.21
16 AUSTRALIA 48.15
17 UKRAINE 43.76
18 SAUDI ARABIA 42.09
19 COSTA RICA 40.89
20 ECUADOR 39.44
21 TUNISIA 36.79
22 CZECH REPUBLIC 35.81
23 JAPAN 35.63
24 SWITZERLAND 33.49
25 SERBIA AND MONTENEGRO 23.11
26 IRAN 20.97
27 GHANA 17.78
28 IVORY COAST 15.20
29 TRINIDAD AND TOBAGO 13.68
30 POLAND 10.09
31 ANGOLA 8.16
32 TOGO 7.33

24
Chapter 2: Ranking Algorithms Andrei C. Grecu

2.2 Random Walkers Rankings

The Intuition

Soccer fans are renowned for the fanatic love and loyalty to their team. However,

for the sake of this algorithm, we need to consider a set of N perpetually vacillating

soccer supporters. Each of these N supporters gets a single vote to designate his or her

favorite team out of the 32 teams qualified for the soccer World Cup. However, over time

each supporter changes her favorite team according to the following rule:

1. She recalls the win-loss outcome of a single game played by her favorite team

2. She flips a weighted coin that is more likely to come up Heads

3. Completely ignoring her current favorite, she goes with the winner of the

game if Heads, and with the loser if Tails

4. Returns to step 1

By constantly looking at historical games between teams and flipping weighted

coins to decide which team to vote for, supporters will perpetually change their mind

regarding their favorite team. Indeed, at the microscopic level voters act as perpetual

random walkers on a network whose nodes are defined by soccer teams, and whose

connections are defined by previous games between the teams. However, even though

each individual voter is a random walker, at a macroscopic level the total number of votes

cast for each team stabilizes over time. Thereafter, we are able to rank the teams

according to fraction of soccer supporters who vote for each team.

The main appeal of this ranking scheme is its simplicity. The only subjective

input for the algorithm is the weight of the coin flipped by the soccer fans- the probability

25
Chapter 2: Ranking Algorithms Andrei C. Grecu

with which they go with the winner of the game they consider. In order to make the

algorithm reasonable, this probability must lie within the interval (0.5, 1), so that the

supporter’s vote goes for the winner, rather than with the loser of the game she consider.

In particular, a coin weight close to the p= 0.5 limit does not guarantee a vote in case of a

victory, but rather rewards the strength of schedule of each team. On the other hand, a

coin weight close to the p= 1 limit almost guarantees a vote in case of a victory, therefore

favoring teams with an undefeated record. With these considerations in mind, I will let

voters in my model go with the winner of the game with the mid-value probability p=

0.75, so that the algorithm equally emphasizes the strength of schedule and the winning

record of the teams.

The Mathematical Model

Consider a competition in which N teams play an uneven paired schedule,

meaning that not all teams play each other. For each team i, let ni be the total number of

games played, wi the number of wins, and li the number of losses. Because soccer allows

tied games, we treat a tie as half a win and half a loss, so that ni = wi + li stays true. Also,

consider V voters, with vi voters casting their single vote to team i, so that even though

voters change their preferences, the total number of voters remains constant, v
i
i V .

At the beginning of the algorithm each voter is randomly assigned a favorite team

out of the N teams to be ranked. Then each of the N voters randomly picks a previous

game played by his favorite team and, completely ignoring his current team preference,

he casts his vote to the winner of that game with probability p, and goes with the loser

with probability 1- p.

26
Chapter 2: Ranking Algorithms Andrei C. Grecu

Additionally, in order to put more emphasis on recent games, I make the

probability of choosing a given game inversely proportional to the number of years since

1
it took place. In order to do this I assign each game the value , where Yk is the number
Yk

of years since the game took place, and I recalculate the weighted sums for the number of

games, wins, and losses for each team. For example, assuming that team i played n

games, won w of them, lost l of them, and tied in t of them, I calculate the discounted

number of games, ni, wins, wi, and loses, li, for team i by summing the discounted value

of each separate game and by treating ties as half-wins and half-losses, so that I maintain

the equality ni = wi + li:

n
1
ni  
k 1 Yk

w t
1 1
wi    0.5* 
k 1 Yk k 1 Yk

l t
1 1
li    0.5* 
k 1 Yk k 1 Yk

Where:

 Yk is the number of years since game k took place


 n i , w i , and li are the discounted number of games, wins, and losses
 n, w, and l are the real numer of games, wins, and losses for team i


In order to avoid rewarding teams for playing more games, I set the rate at which

a voter considers a game played by his favored team i to be independent of the number ni

of games played by the team. With this choice of constant rates, the expected rate of

27
Chapter 2: Ranking Algorithms Andrei C. Grecu

change for the number of votes cast for each team in the random walk is characterized by

the following homogenous system of linear differential equations:

V’ = G * V

Where:

V is the column vector of the number of votes vi cast for each of the N teams


G is a square matrix whose entries are derived from previous games such that:


Gii   p * li  (1  p ) * wi

2* p  1
Gij  0.5* Nij  * Aij , i  j
2
Where:

 Nij is the number of games played between i and j



 A is the number of times i beat j minus the number of times j beat i
 ij

With this choice of matrix G, which encompasses all the outcomes of previous

games between all teams, I am interested in the steady-state equilibrium that gives the

expected number of voters vi who prefer each team:

G*V=O

This equilibrium equation has a unique solution V for any given p in the (0.5,1)

interval. Indeed, with p strictly greater than 0.5, p > 0.5, the off-diagonal elements

2 * p 1
Gij  0.5* N ij  * Aij , i  j of the matrix G are non-negative. If, in addition, the
2

28
Chapter 2: Ranking Algorithms Andrei C. Grecu

underlying graph representing games played between teams consists of a single

connected component, then V is the unique equilibrium solution, lies in the null-space of

G and therefore is the eigenvector associated with the zero eigenvalue. Even though the

equilibrium V does not imply constant flows of voters switching their preference from

team i to team j, it guarantees a constant number of random walkers voting for each team

at any time. Therefore, by reducing the matrix G to its row-reduced echelon form, we can

easily compute the expected number of voters supporting each team, and implicitly rank

the teams according to the overall percentage of votes received.

Application to the 2006 Soccer World Cup

In order to rank the teams that qualified for the 2006 soccer World Cup, I applied

the algorithm described above to the 477 games among the 55 national teams that took

part in at least one World Cup between 1994 and 2006. Using Excel to compute the

discounted number of games, wins, and loses for each team, I constructed the matrix G

used to find the steady-state equilibrium. Then using MATLAB to find the row-reduced

echelon form of G, I determined the relative strength of each of the 55 teams considered.

Choosing the 32 teams qualified for the 2006 World Cup out of the 55 teams, and scaling

the results on a 100-point scale, I obtained the following ranking of teams:

29
Chapter 2: Ranking Algorithms Andrei C. Grecu

2006 WORLD CUP PARTICIPANTS

1 CZECH REPUBLIC 100.00

2 FRANCE 99.97
3 BRAZIL 98.44
4 PORTUGAL 88.17
5 ENGLAND 82.51
6 GERMANY 82.05
7 ITALY 80.78
8 ARGENTINA 80.61
9 NETHERLANDS 77.64
10 SPAIN 77.45
11 MEXICO 72.23
12 SWEDEN 69.22
13 AUSTRALIA 68.97
14 CROATIA 64.09
15 PARAGUAY 60.73
16 USA 60.46
17 SERBIA AND MONTENEGRO 59.99
18 UKRAINE 56.60
19 JAPAN 53.03
20 IRAN 52.84
21 SOUTH KOREA 48.27
22 TUNISIA 47.59
23 POLAND 45.40
24 GHANA 41.04
25 SWITZERLAND 40.50
26 COSTA RICA 38.50
27 IVORY COAST 37.10
28 TOGO 35.53
29 ECUADOR 34.25
30 SAUDI ARABIA 33.17
31 TRINIDAD AND TOBAGO 29.12
32 ANGOLA 27.08

30
Chapter 2: Ranking Algorithms Andrei C. Grecu

2.3 Neural Network Rankings

Artificial neutral networks are biologically inspired algorithms whose pattern-

matching and learning capabilities address problems otherwise difficult to solve using

standard computational and statistical methods. By mathematically modeling the

processes that take place in the human brain, artificial neural networks “learn” from a

given data set, thus being able to develop a solution without needing a predefined

solution algorithm. However, despite their ability to explain complex relationships

between inputs and outputs, artificial neural networks are based on a simple model of

biological neural networks, which makes them intuitively appealing.

Technically speaking, an artificial neural network is characterized by three

elements: i) its processing elements analogous to the neurons in the neural networks, ii)

its connections among processing elements and their corresponding weights, which

contain the information stored in the network, and iii) its evolution algorithm, which

gives the manner in which the neural network searches and converges to a solution.

In an attempt to model the way in which the human brain processes visual data

and learns to recognize objects, psychologist Frank Rosenblatt designed the first artificial

neural network in 1958. Since then, artificial neural network have been applied in a

variety of applications, including bonds rating, target marketing, evaluating loan

applications, predicting stock movements, and certifying signatures. Another application

of neural networks is the rating of sport teams based on previous results, which I will

discuss in this chapter.

31
Chapter 2: Ranking Algorithms Andrei C. Grecu

The Intuition

Wilson (1995) discusses the use of a neural network to rank the NCAA Division

1-A college football teams. In his approach, the processing elements correspond to the

football teams to be ranked, with the value of each neuron representing the relative value

of one team. Also, connections between processing elements correspond to games played

between teams, with the weights of the connections being derived from the score of

previous encounters. Finally, the evolution of the neural network is characterized by the

transfer and summation functions: On one hand, the transfer function look at one game at

a time, takes as inputs the value of the opponent and the weight of the connection

between the two teams, and returns an updated value of the team as a result of that game.

On the other hand, the summation function averages the results of the transfer function

for each game, thus giving the updated value of the team. Eventually, the algorithm stops

when the updated values of all neurons stabilize over time, when a new iteration does not

change the relative value of each neuron.

Using the same setting, I look at the soccer World Cup and design an artificial

neural network that ranks the qualified teams based on their previous results. In order to

do this, I will change the approach presented in Wilson (1995) in two important ways.

First, in assigning weights to the connections corresponding to previous games, I will

factor in both the score and the date of the game, thus putting decreasing emphasis on

games that took place longer time ago. Second, I will implement a soccer-intuitive

transfer function, which updates the values of processing elements (teams) based on the

weight of their connection (score and date of previous game).

32
Chapter 2: Ranking Algorithms Andrei C. Grecu

Consider the following history of games between soccer national teams, in which

the arrow from GER to USA means that team GER beat team USA, and WGER,USA is a

weight that increases with the difference in goals scored by GER and USA, and decreases

with the number of years since the game took place:

GER WGER,USA USA

WSPA,GER WUSA,JAP

WUSA,SPA WGER,JAP

SPA JAP

Figure 3: Neural network setting with 4 neurons representing

teams, and 5 connections representing previous games

Taking the processing elements and connections as inputs, an evolution algorithm

updates the value assigned to each team at time t based on the team values at time t-1, as

well as the connection weights, which remain constant throughout the algorithm. For the

situation pictured in Figure 3, the following updates are performed at each iteration, until

the value assigned to each team does not change between incremental units of time:

33
Chapter 2: Ranking Algorithms Andrei C. Grecu

FT (Vt  1, USA, WGER , USA)  FT (Vt  1, JAP, WGER , JAP)  FT (Vt  1, SPA, WSPA, GER )
Vt , GER 
3

FT (Vt  1, GER, WGER, USA)  FT (Vt  1, JAP, WUSA, JAP)  FT (Vt  1, SPA, WUSA, SPA)
Vt , USA 
3

FT (Vt  1, GER, WSPA, GER )  FT (Vt  1, USA, WUSA, SPA)

Vt , SPA 
2

FT (Vt  1, GER, WGER , JAP)  FT (Vt  1, USA, WUSA, JAP)

Vt , JAP 
2

Where:

 FT is the transfer function with inputs Vt-1 and W



 Vt,ABC is the value of team ABC at time t


 WABC,XYZ is the weight of the connection between team ABC and team XYZ

Intuitively, the transfer function FT should i) increase the value of a team

following a victory, ii) assign more value to a victory against a strong opponent than

against a weaker opponent, iii) assign more value to a recent victory at a high score than

to a close victory a while ago; iv) decrease the value of a team following a defeat, v)

decrease more value following a defeat against a weak opponent than against a stronger

opponent, and finally vi) decrease more value for a recent defeat at a high score than for a

close defeat a while ago. Also intuitively, the initial value assigned to each team should

not influence the final rankings. Thus, in order to avoid any bias resulting from the initial

value assigned to each team, I assign all teams the same initial value and let the

connection weights be the only factors determining the relative ranking of teams. With

34
Chapter 2: Ranking Algorithms Andrei C. Grecu

these considerations in mind, I implement the following mathematical model for a neural

network that ranks the teams qualified for the World Cup.

The Mathematical Model

Consider N teams with a history of games among themselves, such that every

team is connected to all other teams either through direct games or through a succession

of games involving its direct opponents. We consider a corresponding artificial neural

network with N neurons, such that each team is assigned a neuron in the network. Using

the results of previous encounters among the N teams over the last 12 years, we compute

the weights of all connections between team i and team j, such that clear, recent results

are weighted more heavily than tight games, which took place longer ago:

Wij  gij * Yij

Where:

 g ij is the difference between the number of goals scored by team i and team j


 Y is a year coefficient equal to (12 - number of years since the game took place)
 ij

Given the artificial neural network and the weights described above, we need to

assign each team an initial value, before starting the evolution algorithm. As discussed in

the previous section, we want the relative values of the teams to be derived solely from

the connection weights, rather than being influenced by the initial values assigned to the

teams. Therefore, in order to reduce the bias resulting from arbitrary initial values, we

35
Chapter 2: Ranking Algorithms Andrei C. Grecu

assign all teams an initial value of 50, and we implement an evolution algorithm that uses

the connection weights as inputs to rank the teams on a 0 to 100 scale.

At the heart of the evolution algorithm is the transfer function, which updates the

value of each team based on the value of the opponent and the weight of the connection

between the two teams. Intuitively, we want the transfer function to always increase the

value of a team following a victory and decrease its value following a defeat. However,

we want these changes to vary according to the strength of the opponent, the magnitude

of the victory, and the number of years since the game took place. An intuitive transfer

function that accomplishes the objectives discussed above is presented below:

0 Vi,t maxVi,t+1’ maxVi,t+1 100

Figure 4: The value of team i can increase more following a victory against a strong

opponent than a victory against a weaker opponent, maxVi,t+1 > maxVi,t+1’

Consider the case in which team i beat team j, so that the value of team i rises

from Vi,t, at time t, to Vi,t+1, at time t+1. At this point we model a transfer function that

performs two steps in updating the value of team i: i) first, it computes the range of

possible updated values of team i, (minVi,t+1, maxVi,t+1), as a function of the opponent’s

value Vj,t, and then ii) it determines the exact updated value Vi,t+1  (minVi,t+1, maxVi,t+1),

by looking at the connection weight Wij. Such a transfer function generates the greatest

increase in a team’s value following a recent, clear victory against a strong opponent, and

36
Chapter 2: Ranking Algorithms Andrei C. Grecu

causes the smallest increase in value following a tight victory against a weak opponent,

twelve years ago.

First, in determining (minVi, t+1, maxVi, t+1), assume that minVi, t+1 = Vi, t, or the

“worst” a team can do in case of a victory is keep its previous value. Meanwhile, assume

that maxVi,t+1, or the maximum possible increase in the value of team i, is linearly

dependent on the value of the opponent team, Vj,t. Since the team values are bound to the

interval [0,100], consider the two limit cases. On one hand, if team i wins against a 100-

value team, its updated value Vi, t+1 should be allowed to potentially increase all the way

3 3
to of the distance between Vi, t and 100, namely to the point Vi + * (100 – Vi). On
4 4

the other hand, if team i wins against a 0-value team, regardless of the score and date, its

1
update value Vi, t+1 should not be allowed to increase by more than only of the distance
4

1
between Vi, t and 100, or the point Vi + * (100 – Vi). Using these two values as
4

reference, and assuming a linear relationship between the maximum possible increase in

the value of team i, maxVi,t+1, and the value of team j, Vj,t, we derive the formula for the

maximum possible updated value of team i, for a given Vj,t:

100  Vi ,t 100  3* Vi ,t
maxVivictory
,t 1  * V j ,t 
200 4

Second, once we compute (minVi,t+1, maxVi,t+1), or the interval of potential

updated values for team i, we look at the connection weight Wij to determine the exact

updated value for team i within the interval. For large values of Wij, corresponding to

recent, categorical victories, we want the transfer function to assign a big increase in the

37
Chapter 2: Ranking Algorithms Andrei C. Grecu

value of team i. Conversely, for small values of Wij, corresponding to close victories that

took place a while ago, we want a smaller increase in the value of team i within the

interval (minVi,t+1, maxVi,t+1). Therefore, assume that the transfer function returns the

maximum updated value for team i, maxVi,t+1, following a game with weight Wmax, the

highest weight in the entire neural network. For all other connection weights smaller than

Wmax, the transfer function returns an update value for team i in the interval (minVi,t+1,

maxVi,t+1), according to the following formula:

Wij
Vivictory
,t 1  Vi , t  ,t 1  Vi , t )
*(maxVivictory
Wmax

Similarly, in a game in which team i lost to team j we want the value of team i to

decrease as a function of both the value of the opponent team Vj,t and the connection

weight Wij. Now, the “best” a loosing team can do is keep its current value, maxVi, t+1 =

Vi,t, in case it lost in a tight game, against a very strong team, and a long time ago. In the

general case, following a defeat, the updated value of team i will lie in the interval whose

upper limit is the previous value of the team, Vi,i, and whose lower limit is a linear

function of the opponent’s value:

Vi ,t Vi ,t
,t 1 
min Videfeat * V j ,t 
200 4

Having computed the interval of possible updated values for team i, (minVi,t+1,

maxVi,t+1), we look at the connection weight Wij to determine the exact updated value for

team i within the interval. For large values of Wij, corresponding to recent, categorical

defeats, we want the transfer function to assign a big decrease in the value of team i.

38
Chapter 2: Ranking Algorithms Andrei C. Grecu

Conversely, for small values of Wij, corresponding to close defeats that took place a

while ago, we want a smaller decrease in the value of team i within the interval (minVi,t+1,

maxVi, t+1). Therefore, the transfer function, which returns the exact updated value of

team I following a defeat, is a linearly dependent on the connection weight Wij:

Wij
,t 1  Vi , t 
Videfeat ,t 1  Vi , t )
*(minVidefeat
Wmax

Finally, in case of a tie, we want to penalize the team with the higher value and

reward the team with the lower value, since the former was expected to win, while the

latter produced the surprise. However, since ties are only half-upsets, we only award half

the increase computed in case of a victory, and subtract half the value computed in case

of a defeat. Also, in a tie team i and team j score the same number of goals, so the

connection weight degenerates to Wij = 0. Therefore, I choose a transfer function that

updates the team values as a function of the number or years since the game against team

j, Yij. Assuming that team i has a greater value than team j at time t, and supposing that

we are taking into account games at most 12 years ago, the transfer function increases the

value of team j and decreases the value of team i as follows:

victory
(maxVj,t+1 -2*Vj,t )
V tie
j ,t 1  Vj,t  *Yji
24
defeat
(2 * minVi,t+1 -Vi,t )
V tie
i ,t 1  Vi,t  * Yij
12

Using the transfer functions presented here, the evolution algorithm looks at all

the games played by team i, computes the updated value of team i corresponding to each

39
Chapter 2: Ranking Algorithms Andrei C. Grecu

game played, and then averages the separate values to return the overall update value of

team i. This process in repeated until the values of the teams stabilize between

incremental units of time. Finally, in order to avoid having very large or negative values

assigned to certain teams, we normalize the values assigned to each team after each

iteration, such that the highest ranked team has a value of 100 and all other teams have a

proportional value within the interval (0,100).

Application to the 2006 Soccer World Cup

In order to rank the teams that participate in the 2006 World Cup I formed a

neural network whose neurons correspond to the 55 soccer national teams that qualified

for at least one World Cup between 1994 and 2006. Next, I determined the weights of the

intra-neural connections by examining 477 previous games among the 55 teams, played

either in previous World Cups or in competitive regional tournaments that replicate the

World Cup environment- neutral field, strong motivation. Using the C++ program

included in Appendix A, I implemented the transfer functions explained in the previous

section to update the value of each team based on the value of the opponent and the

connection weight corresponding to each game, I obtained the convergent solution for the

value of each team after 178 iterations2:

2
Included in Appendix B is a graph that shows how the algorithm converges to stable values for each team

40
Chapter 2: Ranking Algorithms Andrei C. Grecu

WORLD CUP 2006 PARTICIPANTS

1 BRAZIL 100.00
2 ITALY 93.32
3 FRANCE 85.65
4 ARGENTINA 84.22
5 ENGLAND 82.09
6 GERMANY 81.55
7 PORTUGAL 80.24
8 NETHERLAND 79.88
9 SPAIN 79.47
10 SWEDEN 77.87
11 CZECH REPUBLIC 77.66
12 MEXIC 76.20
13 AUSTRALIA 69.20
14 PARAGUAY 64.69
15 JAPAN 64.23
16 IRAN 63.48
17 IVORY COAST 61.98
18 YUGOSLAVIA 61.19
19 POLAND 61.08
20 TOGO 55.92
21 CROATIA 53.64
22 SOUTH KOREA 52.17
23 COSTA RICA 51.40
24 UKRAINE 50.18
25 GHANA 46.83
26 USA 45.23
27 TUNISIA 44.96
28 SAUDI ARABIA 43.93
29 TRINIDAD AND TOBAGO 28.58
30 SWITZERLAND 25.51
31 ECUADOR 19.72
32 ANGOLA 17.39

41
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

Chapter 3: A Poisson Model for World Cup Games

Chapter 2 presents three algorithms for ranking soccer national teams qualified

for the World Cup, based on historical games among the teams prior to the start of the

competition. Next, this chapter assesses the predictive power of the rankings developed in

the previous chapter by looking at the relationship between initial rankings and game

results in the case of the 2002 World Cup in South Korea and Japan. Thus, in order to

assess the predicting accuracy of the rankings, I will employ the ranking algorithms from

Chapter 2 using historical games available at the beginning of the 2002 World Cup.

Using the 2002 rankings, I will then look at the games from the 2002 World Cup and

examine how the value assigned to each team predicts the number of goals scored by that

team during the tournament. Implicitly, for every game I calculate the probability of

winning assigned to each team and I compare my predictions with the actual results from

South Korea and Japan. Finally, I explain how the structure of games from the World

Cup makes it difficult to compute conditional probabilities for each team to win entire

tournament, which is why I choose to run the simulations of the World Cup in Chapter 4.

Rankings at the start of the 2002 World Cup

Before looking for a probability distribution for the number the goals scored in

the 2002 World Cup games, we need to determine the relative rankings of the teams as of

June 2002, the start date for the tournament. In order to rank the teams at the start of the

World Cup, we apply the algorithms discussed in Chapter 2 to a history of games played

42
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

among soccer national teams between 1990 and 2002. Thus, i) we put decreasing

emphasis on games that took place longer time ago, ii) we only consider competitive

soccer national teams that qualified for at least one World Cup between 1990 and 2002,

and iii) we only look at games that were part of a competitive tournament, in which teams

play their best players at full potential. More precisely, the dataset consists of 473 games

played between 55 teams at the following tournaments:

1. World Cup: France 1998, USA 1994, Italy 1990

2. African Nations Cup: Mali 2002, Ghana 2000, Burkina Faso 1998, South

Africa 1996, Tunisia 1994, Senegal 1992, Algeria 1990

3. Asian Cup: Lebanon 2000, United Arab Emirates 1996, Japan 1992

4. CONCACAF Gold Cup: USA 2002, USA 2000, USA 1998, USA 1996,

Mexico 1993, USA 1991

5. Copa America: Colombia 2001, Paraguay 1999, Bolivia 1997, Uruguay 1995,

Ecuador 1993, Chile 1991

6. European Championship: Belgium/Netherlands 2000, England 1996, Sweden

1992

By applying the three ranking algorithms to this set of games, with their

corresponding dates and scores, I obtained the following rankings of soccer national

teams at the beginning of the 2002 World Cup:

43
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

2002 WORLD CUP EIGENVECTOR RANKING

1 ITALY 100.00
2 ARGENTINA 95.66
3 GERMANY 92.36
4 BRAZIL 90.81
5 MEXICO 82.73
6 USA 76.02
7 ENGLAND 74.46
8 FRANCE 64.60
9 URUGUAY 62.42
10 CAMEROON 61.64
11 SPAIN 61.27
12 SWEDEN 60.79
13 PARAGUAY 54.63
14 BELGIUM 49.56
15 RUSSIA 47.92
16 DENMARK 46.81
17 COSTA RICA 45.92
18 NIGERIA 44.96
19 ECUADOR 44.48
20 SOUTH KOREA 42.81
21 CROATIA 39.88
22 PORTUGAL 34.62
23 IRELAND 31.84
24 SAUDI ARABIA 31.02
25 TUNISIA 28.65
26 SOUTH AFRICA 25.76
27 JAPAN 23.17
28 TURKEY 21.79
29 SENEGAL 12.75
30 SLOVENIA 8.78
31 CHINA 7.78
32 POLAND 7.00

44
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

2002 RANDOM WALKERS RANKING

1 FRANCE 100.00
2 PORTUGAL 82.90
3 ITALY 75.27
4 CROATIA 63.80
5 BRAZIL 59.04
6 GERMANY 54.69
7 ENGLAND 51.22
8 ARGENTINA 48.08
9 SPAIN 46.93
10 IRELAND 45.06
11 MEXIC 42.30
12 DENMARK 40.77
13 NIGERIA 40.63
14 CAMEROON 39.67
15 SENEGAL 39.42
16 USA 36.58
17 BELGIUM 36.44
18 SWEDEN 34.58
19 PARAGUAY 33.33
20 URUGUAY 32.84
21 TURKEY 32.37
22 RUSSIA 32.19
23 SLOVENIA 31.16
24 SOUTH AFRICA 30.81
25 JAPAN 26.20
26 ECUADOR 22.84
27 COSTA RICA 22.60
28 TUNISIA 21.40
29 SAUDI ARABIA 20.82
30 SOUTH KOREA 19.13
31 CHINA 10.53
32 POLAND 10.00

45
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

2002 NEURAL NETWORK RANKING

1 ITALY 100.00
2 PORTUGAL 96.20
3 FRANCE 94.56
4 BRAZIL 84.97
5 CROATIA 82.69
6 DENMARK 47.03
7 ENGLAND 46.11
8 NIGERIA 45.64
9 GERMANY 45.56
10 SPAIN 44.91
11 ARGENTINA 44.91
12 SOUTH AFRICA 37.88
13 CAMEROON 33.92
14 MEXIC 31.44
15 PARAGUAY 30.82
16 SLOVENIA 29.60
17 SWEDEN 28.87
18 TURKEY 26.81
19 URUGUAY 26.37
20 TUNISIA 25.26
21 SENEGAL 24.97
22 BELGIUM 23.87
23 JAPAN 23.41
24 RUSSIA 19.60
25 USA 19.56
26 IRELAND 19.16
27 ECUADOR 19.05
28 COSTA RICA 18.74
29 SOUTH KOREA 18.33
30 SAUDI ARABIA 18.22
31 CHINA 2.57
32 POLAND 2.00

46
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

A Poisson Distribution of Goals

In a typical soccer game two teams play in continuous time for 90 minutes3, with

the team that scores more goals winning the game. In order to score a goal, the two teams

alternate possession of the ball, which allows them to attack the opponent team’s goal

and perhaps score. If attacks are assumed to be independent, then the distribution of goals

scored is negative binomial, as claimed in Pollard (1977), Moroney (1951), and Reed,

Pollard, and Benjamin (1971). However, the probability p that any attack will result in a

goal is relatively small. For example, a total of 161 goals were scored in the 64 games

from the 2002 World Cup, for an average of approximately 2.5 goals per game or one

goal every 40 minutes. Consequently, even thought there are a lot of attacks in a game,

there is a small probability p of scoring in any given attack. Therefore, given the small

probabilities of success p, the Poisson approximation is a good fit for the number of goals

scored in a soccer game, as showed in Lee (1997), Maher (1982), and Dyte and Clarke

(2000).

Using an approach similar to Dyte and Clarke (2000), my model for the number

of goals scored by each team in a given game relies on two assumptions. First, it assumes

that the number of goals scored by each team in a game is Poisson distributed. Second, it

assumes that the number of goals scored by a team is independent of the number of goals

scored by the opposing team. Given these two assumptions, I am interested in measuring

the rankings’ predictive power with respect to the number of goals scored by team i in a

game against team j. In particular, I am looking for the following relationship:

3
59 out of the 64 games played in the 2002 World Cup ended after 90 minutes. Thus, even though some
games from the knockout stage of the World Cup go into extra time to allow teams to score the winning
goal, I will focus my analysis on results after 90 minutes of play.

47
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

ln( gij )  x0  x1 * Ri  x2 * R j

Where:

 gij is the number of goals scored by team i in a game against team j



 R is the rating of team i at the time of the game
 i

Having agreed on the model, I looked at the results from the games played in the

2002 World Cup. Even though only 64 games were played in the tournament, they

correspond to 128 observations, since for every game we are interested in predicting the

number of goals scored by each team. Using MATLAB to fit a Poisson regression to the

data, I obtained the following sets of coefficients, corresponding to each of the three

rankings developed:

1. Eigenvector Ranking:

Value t value p value

X0 0.3532 1.5902 0.1118
X1 0.0071 2.4309 0.0151
X2 -0.0106 -3.4533 0.0006
Deviance 140.9387
Dispersion 1.0134

2. Random Walkers Ranking:

Value t value p value

X0 0.6224 2.279 0.0227
X1 0.0046 1.1241 0.261
X2 -0.0153 -3.0812 0.0021
Deviance 145.7578
Dispersion 1.0237

48
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

3. Neural Network Ranking:

Value t value p value

X0 0.3976 1.9985 0.0457
X1 0.0039 1.3371 0.1812
X2 -0.0091 -2.4773 0.0132
Deviance 148.7348
Dispersion 1.0429

All three regressions fit the data well, and give a positive coefficient in front of Ri,

the rating of team i, and negative coefficient in front of Rj, the rating of the opponent

team j. These results confirm the basic intuition, which says that the higher the rating of

team i, and the lower the ranking of team j, the higher the expected number of goals

scored by team i. Also, the negative coefficients in front of Rj are significantly larger in

absolute terms than the positive coefficients in front of Ri for all three regressions,

suggesting that in determining the number of goals scored by a team, the rating of the

opponent is more important than the team’s own ranking, at least in the 2002 World Cup.

Using the regression coefficients to compute the expected number of goals scored

by a team in a certain game and then using that expectation as a mean, we can compute

marginal probabilities for each team’s Poisson distribution of goals scored in a certain

game. If we further assume that the number of goals scored by team i is independent of

the number of goals scored by team j, we can compute the probability of any specific

result for a given game:

i  exp( x0  x1 * Ri  x2 * R j )

 j  exp( x0  x1 * R j  x2 * Ri )

49
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

g  j
i ij * e i  j ji * e
g

P( gij & g ji )  *
gij ! g ji !

As an example, consider the game between Argentina and Nigeria from the 2002

World Cup. The Eigenvector ratings computed at the beginning of the tournament for the

two teams are RARG = 95.66 and RNIG = 61.64. Using these ratings, we compute the

expected number of goals scored by each team in their encounter:

GER  exp( x0  x1 * RARG  x2 * RNIG ) 

 exp(0.3532  0.0071*95.66  0.0106*61.64) 
 1.46

NIG  exp( x0  x1 * RNIG  x2 * RARG ) 

 exp(0.3532  0.0071*61.64  0.0106*95.66) 
 0.79

Thus, before the game we expected Argentina to score 1.46 goals and Nigeria

only 0.79 goals in their direct game. Using these expectations as means, we can compute

the probability assigned to each score in the game between Argentina and Nigeria, as

long as we assume that the number of goals scored by each team is Poisson distributed

and the number of goals scored by Argentina is independent of the number of goals

scored by Nigeria:

ARG
g ARG
* e  ARG NIG
g NIG
* e  NIG
P (1-0 Argentina win)  * 
g ARG ! g NIG !
 1.46* e 1.46 * e 0.79 
 0.149

This is the probability of a 1 - 0 Argentina victory over Nigeria. Performing

similar computations, we can compute the probability assigned to any score in the game

50
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

between Argentina and Nigeria. The table below contains the probabilities assigned to

any result in which both teams score between 0 and 9 goals:

NIG
ARG 0 1 2 3 4 5 6 7 8 9
0 0.0853 0.0607 0.0216 0.0051 0.0009 0.0001 0.0000 0.0000 0.0000 0.0000
1 0.1493 0.1062 0.0378 0.0090 0.0016 0.0002 0.0000 0.0000 0.0000 0.0000
2 0.1306 0.0929 0.0331 0.0078 0.0014 0.0002 0.0000 0.0000 0.0000 0.0000
3 0.0762 0.0542 0.0193 0.0046 0.0008 0.0001 0.0000 0.0000 0.0000 0.0000
4 0.0333 0.0237 0.0084 0.0020 0.0004 0.0001 0.0000 0.0000 0.0000 0.0000
5 0.0117 0.0083 0.0030 0.0007 0.0001 0.0000 0.0000 0.0000 0.0000 0.0000
6 0.0034 0.0024 0.0009 0.0002 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
7 0.0009 0.0006 0.0002 0.0001 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
8 0.0002 0.0001 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
9 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

Table 1: Probability distribution of scores for Argentina vs. Nigeria

In this example, it happens to be that the score with the highest probability

(14.93%), 1-0 for Argentina, was the real result of the game between the two teams.

However, given the relatively small probability assigned to each separate result, it does

not happen very often that the result with the highest probability is the real result of the

game. A more meaningful prediction regarding the game between Argentina and Nigeria

would be the probability assigned to a victory of Argentina against Nigeria, at any score.

We are able to compute this probability by summing all the probabilities below the main

diagonal- the probabilities assigned to all outcomes in which Argentina scores at least

one goal more than Nigeria. Similarly, we compute the probability of a Nigeria victory by

summing the probabilities above the main diagonal. Finally, the sum of the probabilities

on the main diagonal gives the probability of a draw. In the case of the game between

51
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

Argentina and Nigeria, these aggregate probabilities were: Argentina wins 62.30%,

Nigeria wins 14.75%, game ends in a tie 22.95%.

Predictive Accuracy for the 2002 World Cup Games

Using the methodology discussed above, I compute three sets of predictions for

the 2002 World Cup games, each corresponding to one of the three rankings presented in

this thesis: the Eigenvector ranking, the Random Walker ranking, and the Neural

Network ranking. Then, by looking at the actual results of the games from the 2002

tournament I check whether my predictions turned out to be correct or not. However,

given the relative frequency of unexpected results in the soccer World Cup, any initial

ranking will inherently generate a relatively large number of wrong predictions.

Therefore, a more meaningful assessment of the predictive power of my rankings would

be a comparison with the predictions of another party interested in correctly forecasting

the outcome of World Cup games. Naturally, I decided to compare my predictions with

the predictions implied by the odds for winning the World Cup assigned to each team by

the sports bookies. More exactly, I assumed that whenever a team with higher odds for

the World Cup plays a team with lower such odds, the former team is the bookies’

favorite.

Below are the odds for winning the 2002 World Cup assigned to each team by the

London bookmakers, as published in The London Times on the 31st of May, 2002, the

day before the start of the tournament:

52
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

THE LONDON TIMES RANKING

$ FOR $1 BET P (WIN THE WC)

1 FRANCE 3.33 20.54%
2 ARGENTINA 4.5 15.20%
3 ITALY 4.5 15.20%
4 BRAZIL 6 11.40%
5 SPAIN 8 8.55%
6 ENGLAND 12 5.70%
7 PORTUGAL 14 4.89%
8 GERMANY 16 4.28%
9 CAMEROON 33 2.07%
10 RUSSIA 66 1.04%
11 JAPAN 66 1.04%
12 PARAGUAY 80 0.86%
13 SWEDE 80 0.86%
14 CROATIA 80 0.86%
15 POLAND 80 0.86%
16 NIGERIA 100 0.68%
17 URUGUAY 100 0.68%
18 TURKEY 100 0.68%
19 MEXIC 125 0.55%
20 BELGIUM 125 0.55%
21 IRELAND 125 0.55%
22 DENMARK 125 0.55%
23 ECUADOR 150 0.46%
24 SOUTH KOREA 150 0.46%
25 SENEGAL 200 0.34%
26 SLOVENIA 300 0.23%
27 USA 300 0.23%
28 SOUTH AFRICA 300 0.23%
29 COSTA RICA 400 0.17%
30 TUNISIA 500 0.14%
31 SAUDI ARABIA 750 0.09%
32 CHINA 750 0.09%

53
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

In order to compare my predictions with the London Times predictions, I look at

the 64 games from the 2002 World Cup and I reward the predictions that turned out to be

correct, while I penalize the predictions that turned out to be wrong. Specifically, I assign

a score of –1 to any game in which a team with smaller winning probability defeated a

team with higher winning probability, since this disagrees with the predictions. Likewise,

I assign a score of +1 to any game won by the team with higher winning probability,

since this agrees with the predictions. Intuitively, a tie is worth 0, since the team with

higher winning probability failed to win the game, but it did not actually lose to the

underdog, either.

Also, we define the predictive score of a set of predictions as the sum of the

separate scores assigned to each game, given that set of predictions. The higher the

predictive score, the more accurate the predictions are, since we have decided to award a

point for every correct forecast and subtract one point for every erroneous prediction. For

example, using the predictions derived from the London Times rankings and looking at

the 2002 World Cup results, we conclude that the favorite team actually won the game in

32 out of the 64 games, the underdog produced the surprise and won in 18 games, while

14 games ended in a tie. Thus, with 32 correct predictions, 18 wrong ones, and 14 ties,

the London Times set of predictions gets a predictive score equal to 32*1 – 18*(-1) +

14*0 = 14.

Similarly, we compute the predictive score of the three sets of predictions derived

from my rankings. Using the Eigenvector set of predictions we are able to correctly

predict the winner of 35 games, while we are wrong for 15 games, for a predictive score

of 35*1 + 15*(-1) = 20. Using the Random Walkers set of predictions we again correctly

54
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

predict the winner for 35 of the games, while we are wrong in 15 cases, for a predictive

score of 35*1 + 15*(-1) = 20. Finally, with the Neural Network set of predictions we are

right in predicting the winner of 33 games, while we are mistaken in our prediction for 17

games, for an aggregate predictive score of 33*1 + 17*(-1) = 16.

It turns out that my predictions unanimously outperform the London Times

predictions in picking the winner of separate World Cup games. In other words, if I had

chosen the winner of each game according to my rankings, I would have done better than

if I had chosen the winner of each game by looking in the London Times and picking the

team with the higher assigned odds for winning the World Cup.

55
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

Table 2: For each of the 64 games from the 2002 World Cup, the three sets of

predictions derived from each of the three rankings presented in Chapter 2 (a

Score of 1 means the prediction was right , a –1 means it was wrong; 0 is a tie )

PREDICTIONS DERIVED FROM THE RANKINGS:

Eigenvector Random Walkers Neural Network Times

64 WC Games Winner: Winner: Winner:
A B Result A B Tie Score A B Tie Score A B Tie Score Score
FRA SEN 0 1 65% 14% 21% -1 65% 11% 24% -1 62% 15% 23% -1 -1
URU DEN 1 2 44% 28% 27% -1 32% 41% 27% 1 29% 45% 27% 1 -1
FRA URU 1 2 37% 35% 28% -1 70% 9% 21% -1 61% 15% 24% -1 -1
SEN DEN 1 2 20% 56% 24% 1 35% 37% 27% 1 28% 45% 27% 1 1
FRA DEN 1 2 45% 27% 27% -1 64% 11% 24% -1 52% 20% 27% -1 -1
SEN URU 3 3 15% 64% 21% 0 41% 33% 27% 0 36% 37% 26% 0 0
PAR SAF 2 2 52% 23% 25% 0 38% 35% 26% 0 34% 39% 27% 0 0
SPA SLO 3 1 66% 14% 20% 1 46% 28% 27% 1 42% 31% 27% 1 1
SPA PAR 3 1 39% 33% 28% 1 44% 29% 27% 1 42% 31% 27% 1 1
SLO SAF 0 1 28% 47% 25% 1 37% 37% 26% -1 33% 40% 27% 1 -1
SLO PAR 1 3 16% 62% 22% 1 36% 38% 26% 1 36% 37% 27% 1 1
SPA SAF 3 2 56% 20% 24% 1 46% 27% 26% 1 39% 34% 28% 1 1
BRA TUR 2 1 73% 10% 18% 1 52% 22% 26% 1 58% 17% 25% 1 1
CHN CRI 0 2 19% 58% 23% 1 30% 46% 23% 1 31% 44% 25% 1 1
BRA CHN 4 0 80% 6% 14% 1 67% 13% 20% 1 69% 12% 20% 1 1
TUR COS 1 1 25% 50% 25% 0 44% 31% 25% 0 40% 34% 26% 0 0
BRA CRI 5 2 59% 17% 24% 1 58% 18% 24% 1 62% 15% 23% 1 1
TUR CHN 3 0 45% 30% 25% 1 53% 25% 23% 1 48% 28% 25% 1 1
SKO POL 2 0 57% 20% 23% 1 45% 32% 23% 1 44% 31% 25% 1 -1
USA PTG 3 2 58% 18% 24% 1 15% 60% 25% -1 13% 65% 22% -1 -1
POL PTG 0 4 23% 53% 24% 1 7% 78% 15% 1 10% 72% 18% 1 1
SKO USA 1 1 21% 54% 26% 0 27% 49% 24% 0 37% 38% 26% 0 0
POL USA 3 1 9% 74% 17% -1 22% 55% 22% -1 30% 45% 25% -1 1
SKO PTG 1 0 41% 32% 27% 1 9% 72% 19% -1 13% 65% 22% -1 -1
GER SAU 8 0 68% 12% 20% 1 58% 19% 24% 1 48% 27% 26% 1 1
IRE CAM 1 1 22% 53% 25% 0 39% 33% 28% 0 31% 43% 26% 0 0
GER IRE 1 1 68% 12% 20% 0 41% 30% 29% 0 47% 27% 26% 0 0
SAU CAM 0 1 22% 53% 25% 1 26% 49% 25% 1 31% 43% 26% 1 1
GER CAM 2 0 51% 22% 27% 1 44% 28% 28% 1 41% 32% 27% 1 1
SAU IRE 0 3 36% 37% 27% 1 23% 52% 24% 1 37% 37% 26% 1 1
ARG NIG 1 0 62% 15% 23% 1 40% 32% 28% 1 36% 36% 28% -1 1
ENG SWE 1 1 43% 29% 28% 0 46% 27% 27% 0 43% 30% 27% 0 0
ARG ENG 0 1 46% 26% 29% -1 34% 37% 29% 1 36% 36% 28% 1 -1
NIG SWE 1 2 28% 44% 27% 1 40% 33% 27% -1 43% 30% 27% -1 1

56
Chapter 3: A Poisson Model for World Cup Games Andrei C. Grecu

ARG SWE 1 1 53% 20% 26% 0 44% 29% 27% 0 43% 30% 27% 0 0
NIG ENG 0 0 22% 52% 26% 0 30% 42% 28% 0 36% 36% 28% 0 0
CRO MEX 0 1 17% 59% 24% 1 47% 24% 28% -1 55% 19% 26% -1 -1
ITA ECU 2 0 65% 13% 22% 1 66% 12% 21% 1 66% 12% 21% 1 1
ITA CRO 1 2 67% 12% 21% -1 39% 28% 33% -1 40% 28% 32% -1 -1
ECU MEX 1 2 19% 56% 25% 1 26% 49% 25% 1 32% 42% 26% 1 1
ECU CRO 1 0 39% 34% 27% 1 16% 61% 23% -1 16% 61% 23% -1 -1
ITA MEX 1 1 43% 27% 29% 0 53% 19% 28% 0 61% 15% 24% 0 0
JAP BEL 2 2 24% 51% 25% 0 31% 43% 26% 0 37% 37% 26% 0 0
RUS TUN 2 0 47% 27% 26% 1 44% 31% 25% 1 35% 39% 26% -1 1
JAP RUS 1 0 24% 50% 25% -1 34% 41% 26% -1 39% 35% 26% 1 -1
BEL TUN 1 1 48% 26% 26% 0 47% 28% 25% 0 36% 37% 26% 0 0
BEL RUS 3 2 37% 35% 28% 1 39% 34% 26% 1 39% 35% 26% 1 -1
JAP TUN 2 0 34% 40% 26% -1 41% 35% 25% 1 36% 38% 26% -1 1
GER PAR 1 0 55% 19% 26% 1 49% 25% 27% 1 42% 31% 27% 1 1
DEN ENG 0 3 23% 51% 26% 1 30% 42% 28% 1 36% 36% 28% -1 1
SWE SEN 1 2 63% 15% 21% -1 34% 39% 27% 1 38% 35% 26% -1 -1
SPA IRE 1 1 52% 22% 25% 0 37% 35% 29% 0 47% 27% 26% 0 0
MEX USA 0 2 39% 32% 29% -1 40% 33% 27% -1 42% 32% 26% -1 -1
BRA BEL 2 0 57% 18% 25% 1 49% 24% 27% 1 59% 17% 24% 1 1
JAP TUR 0 1 38% 36% 26% -1 33% 41% 26% 1 35% 38% 26% 1 -1
SKO ITA 2 1 13% 66% 22% -1 11% 69% 20% -1 12% 67% 21% -1 -1
ENG BRA 1 2 28% 43% 29% 1 31% 39% 30% 1 23% 50% 28% 1 1
GER USA 1 0 43% 28% 29% 1 46% 26% 27% 1 47% 27% 26% 1 1
SKO SPA 0 0 27% 46% 27% 0 22% 55% 24% 0 27% 47% 26% 0 0
SEN TUR 0 1 32% 42% 26% 1 41% 32% 27% -1 36% 38% 26% 1 1
SKO GER 0 1 15% 62% 23% 1 18% 59% 23% 1 27% 48% 26% 1 1
BRA TUR 1 0 73% 10% 18% 1 52% 22% 26% 1 58% 17% 25% 1 1
SKO TUR 2 3 48% 26% 26% -1 29% 46% 25% 1 34% 40% 26% 1 1
BRA GER 2 0 34% 36% 30% -1 37% 32% 31% 1 50% 22% 28% 1 1
SCORES: 20 20 16 14

57
Chapter 4: Simulating the Soccer World Cup Andrei C. Grecu

Chapter 4: Simulating the Soccer World Cup

Chapter 3 showed how the results of the ranking algorithms from Chapter 2 could

be used to make predictions regarding the outcome of separate games from the World

Cup. Moreover, we have seen that for the 2002 World Cup games, picking winners using

the three ranking methods outperformed the strategy of always going with the

bookmakers’ favorite in each separate game.

In this chapter we are going to shift our focus from separate World Cup games to

the overall picture and determine the probability of winning the Cup assigned to each

participating team. In particular, how is the probability of winning assigned to each team

influenced by the particular schedule of games from the World Cup tournament? How

does the initial draw advantage some teams, while disadvantaging other participants?

Finally, which teams had a lucky draw for the World Cup final tournament and which

teams were assigned the most difficult schedule? Given the structure of games in the

World Cup, it turns out to be easier to answer these questions by performing a simulation

of the games from the tournament than by explicitly calculating conditional probabilities

of certain events happening.

The Structure of the World Cup Tournament

The soccer World Cup has a complex, two-phase structure. Initially, the 32

qualified teams are divided into eight groups of four. Even though this allocation is based

upon a draw, teams are also divided into four value-pools according to their previous

58
Chapter 4: Simulating the Soccer World Cup Andrei C. Grecu

international results, and each group is restricted to exactly one team from each value-

pool. Theoretically, this structure advantages the highest ranked teams, which are seeded

such that they play lower ranked teams in the group stage. However, teams that

participate in the World Cup are of comparable value and during the six round robin

games played in each group, teams have three games to demonstrate their superiority

over the other teams from the group. At the end of the group games, the first criterion for

ranking the teams is the number of points achieved by each team in the three group

games; in case of a tie, the next criterion used to differentiate between teams is the goal

difference record of each team.4Given the final ranking, the top two teams from each

group advance to the second round of the competition, which consists of a simple 16-

team knockout arrangement. Moreover, the entire structure of games of the knockout

stage is predetermined, such that the winner of group A plays the runner up of group B,

the runner up of group C plays the winner of group D, and so on all the way to the final.

This particular structure essentially splits the teams in two different halves, such that

teams from one half will not play teams from the other half until the final of the World

Cup, assuming that both teams reach that phase.

The tree structure that describes the schedule of games from the knockout stage of

the World Cup tournament is depicted below. Since 48 games are played in the group

stage, the figure labels games G49 to G62, leading all the way to the final:

4
It very rarely happens that these two criteria are not sufficient for ranking the teams, but other criteria for
differentiating exist, such as the number of red/yellow cards received by each team, or the official
international ranking of the teams at the time of the play.

59
Chapter 4: Simulating the Soccer World Cup Andrei C. Grecu

WORLD CUP GAME SCHEDULE

Last 16 Quarter-Finals Semi-finals FINAL Semi-finals Quarter-Finals Last 16

WA G49 G51 WB
W49 W51
RB RA
G57 G59
W57 W59
WD
WC G50 G52
W50 W62 G62 W52 RC
RD

FINAL
WE WF
G53 G61 G55
W53 W61 W55
RF RE
G58 G60
W58 W60

WG WH
G54 G56
W54 W56
RH RG

Figure 5: The tree structure describing the games from the World Cup tournament. E.g.:

The winner of group A (WA) plays the runner-up of group B (RB) in game G49

In theory, we could explicitly calculate conditional probabilities of winning the

Cup for each participating team. However, even if we calculated conditional probabilities

involving winning probabilities in games against all other 31 teams from the World Cup,

we would still run into difficulties when ranking teams after the group stage games, since

goal difference required for tie-breaking make the exact score a requirement. Therefore, I

choose to simulate the games from the World Cup using a method inspired from

Winston, and Albright (2001).

60
Chapter 4: Simulating the Soccer World Cup Andrei C. Grecu

The Simulation Model

Assume that at the beginning of the World Cup the value of team i is Vi, the

corresponding value from the Eigenvector rankings. This simulation models the outcome

of a game between team i and team j by generating two numbers from two normal

distributions with equal variance σ2: the first number Pi is drawn from the distribution

N(Vi, σ2) and characterizes team i’s performance in the game; the second number Pj is

drawn from the distribution N(Vj, σ2) and characterize team j’s performance in the game.

Given these two numbers generated from N(Vi, σ2) and N(Vj, σ2), the score Sij of the

game between team i and team j is defined as:

Sij  Pi  Pj = N(Vi ,  )  N(Vj ,  )

The equal variance σ of the two distributions determines how often will the

sample from the population with the larger mean be lower than the sample from the

population with the smaller mean. This probability is particularly important for our

interpretation of the model, since a negative score Sij = Pi – Pj < 0 for Vi > Vj translates

into a “surprise”, or a game in which the team with lower value Vj defeated the team with

higher value Vi. Therefore, for all the 48 games from the group stage, I look at the 48

values of the “favorite” team in each game, Vf, and I look at the 48 values of the

“underdog” team in each game, Vu. The Central Limit Theorem tells us that Vf  Vu has

approximately a normal distribution regardless of the underlying population distributions.

Therefore, I compute the mean and variance of the two sample sizes, the “favorites” and

the “underdogs”, and I determine the variance σ2 required to make P (Vf – Vu > 0) = 0.75,

assuming that 25% of the games in the World Cup are upsets.

61
Chapter 4: Simulating the Soccer World Cup Andrei C. Grecu

Sample Size Sample Mean Sample Variance

Favorites 48.00 63.67 380.51
Underdogs 48.00 28.31 295.17

Table 3: Mean and variance for the groups of teams with higher value and lower

value, respectively, before the 48 games from the group stage of the World Cup

Thus for σ2 = 58, this simulation model will produce an “upset”- a situation in

P
which given Vi > Vj, the returned sample P
, with probability 0.25.
i j

Simulating the 2002 World Cup

Using the model and the parameters described above, I simulate the games from

the 2002 World Cup. First, I simulate the 6 games from each group and for each team I

aggregate the three scores corresponding to its games. Then I rank the teams according to

their aggregate scores and, having obtained the top two teams in each group, I simulate

the predefined schedule of games until I get the winner of the World Cup. I repeat the

simulation 10,000 times and I count the number of times each team won the World Cup,

thus determining the probability of winning the World Cup for each team.

Having computed the probability for winning the World Cup for each team, we

want to determine which teams were advantaged, and which teams were disadvantaged

by the World Cup draw. It is generally agreed upon the fact that a Round Robin

tournament, in which each team plays every other team, gives a fair ranking of the teams.

Therefore, I also simulate a Round Robin alternative to the World Cup. I only run the

62
Chapter 4: Simulating the Soccer World Cup Andrei C. Grecu

Round Robin simulation 1000 times, since every tournament takes 992 games, as

opposed to only 64 in the case of the World Cup, and I compute the probability of

winning the Round Robin tournament for each team. Then by comparing the results of

the World Cup simulatiuon with the results of the Round Robin simulation, I can

determine which teams were mostly advantages or disadvantaged by the World Cup

draw. Here are the results of the 2002 World Cup and Round Robin simulations- apart

from Italy, Brazil shows the greatest World Cup advantage, or difference between its

winning probabilities for the World Cup and for the Round Robin tournament, which

might partially explain their successful advancement all the way to the final and to

winning the Cup:

2002 SIMULATION RESULTS:

Teams Round Robin World Cup World Cup Advantage
ITA 6.89% 15.32% 8.43%
ARG 6.69% 12.86% 6.17%
GER 6.34% 9.90% 3.56%
BRA 6.25% 13.44% 7.19%
MEX 5.68% 8.06% 2.38%
USA 5.13% 7.20% 2.07%
ENG 4.97% 4.50% -0.47%
FRA 4.29% 3.94% -0.35%
URU 4.10% 3.62% -0.48%
CAM 4.08% 2.50% -1.58%
SPA 4.01% 3.76% -0.25%
SWE 3.97% 1.44% -2.53%
PAR 3.49% 2.28% -1.21%
BEL 3.22% 1.66% -1.56%
RUS 3.07% 1.40% -1.67%
DEN 2.96% 1.34% -1.62%
CRI 2.86% 1.58% -1.28%
NIG 2.80% 0.64% -2.16%
ECU 2.79% 0.66% -2.13%
SKO 2.63% 1.08% -1.55%
CRO 2.41% 0.46% -1.95%
PTG 1.99% 0.52% -1.47%
IRE 1.76% 0.32% -1.44%

63
Chapter 4: Simulating the Soccer World Cup Andrei C. Grecu

SAU 1.75% 0.24% -1.51%

TUN 1.64% 0.34% -1.30%
SAF 1.34% 0.34% -1.00%
JAP 1.19% 0.18% -1.01%
TUR 1.11% 0.20% -0.91%
SEN 0.41% 0.10% -0.31%
SLO 0.13% 0.04% -0.09%
CHN 0.05% 0.02% -0.03%
POL 0.00% 0.06% 0.06%

Predictions for the 2006 World Cup

Running the simulation for the 2006 World Cup shows Brazil as the big favorite

to win the World Cup, having the highest rating in both tournaments, as well as the

highest World Cup advantage, or increase from the Round Robin to the World Cup

winning percentages, indicating a “lucky draw”:

2002 SIMULATION RESULTS:

World Cup
Teams Round Robin World Cup Advantage
BRAZIL 6.89% 17.58% 10.69%
ITALY 6.69% 11.04% 4.35%
GERMANY 6.34% 10.24% 3.90%
MEXIC 6.25% 8.42% 2.17%
NETHERLAND 5.68% 6.96% 1.28%
ARGTENTINA 5.13% 6.20% 1.07%
FRANCE 4.97% 6.32% 1.35%
ENGLAND 4.29% 4.54% 0.25%
SPAIN 4.10% 2.36% -1.74%
PARAGUAY 4.08% 3.56% -0.52%
SWEDEN 4.01% 4.16% 0.15%
CROATIA 3.97% 2.80% -1.17%
PORTUGAL 3.49% 0.00% -3.49%
USA 3.22% 0.30% -2.92%
SOUTH KOREA 3.07% 0.00% -3.07%
SAUDI ARABIA 2.96% 1.00% -1.96%
COSTA RICA 2.86% 0.00% -2.86%
ECUADOR 2.80% 0.98% -1.82%
TUNISIA 2.79% 0.02% -2.77%
CZECH REPUBLIC 2.63% 0.48% -2.15%

64
Chapter 4: Simulating the Soccer World Cu Andrei C. Grecu

JAPAN 2.41% 0.38% -2.03%

SWITZERLAND 1.99% 3.16% 1.17%
UKRAINE 1.76% 0.94% -0.82%
YUGOSLAVIA 1.75% 2.34% 0.59%
IRAN 1.64% 0.22% -1.42%
AUSTRALIA 1.34% 0.20% -1.14%
GHANA 1.19% 0.10% -1.09%
IVORY COAST 1.11% 0.10% -1.01%
TRINIDAD & TOBAGO 0.41% 0.04% -0.37%
POLAND 0.13% 0.08% -0.05%
ANGOLA 0.05% 0.16% 0.11%
TOGO 0.00% 0.80% 0.80%

65
Chapter 5: Conclusion Andrei C. Grecu

Chapter 5: Conclusion

This thesis proposes a new approach to ranking the teams that qualify for the

soccer World Cup. In essence, it claims that the quality of soccer teams is different on

different continents and therefore results in regional qualifying tournaments, against less

competitive teams, are not so relevant to the present value of the teams. Instead, it looks

at historical games between competitive teams on different continents and implements

three algorithms that take into account the strengths of the opponents encountered in the

past when determining the present value of the teams.

Chapter 2 lays the foundation of the thesis by presenting and then implementing

three algorithms for ranking teams in uneven paired competitions. The three algorithms

use different mathematical tools to rank the teams. First, a matrix-based method of

ranking takes as input previous interactions among teams and returns an eigenvector with

the relative value assigned to each team. Second, a random walker algorithm looks at the

steady-state solution of a setting in which a number of voters continuously change their

mind regarding their favorite team, thus executing random walks on a network defined by

the participating teams and their previous interactions. Third, a biologically inspired

neural network algorithm constructs a neural network whose nodes are the participating

teams and whose connections are determined by previous interactions among the teams,

and uses a soccer-intuitive transfer function to update the value of each node until a

steady state solution is reached. Although very different in their approach, all three

algorithms do a good job in ranking the teams, as tested in Chapter 3.

66
Chapter 5: Conclusion Andrei C. Grecu

Chapter 3 contains the most meaningful results of this thesis. Looking four years

back at the 2002 World Cup in South Korea and Japan, and assuming a Poisson

distribution of goals in a World Cup game, a nonlinear regression for the number of goals

scored in each separate game from the 2002 World Cup is fit, using the rankings

presented above as predictive variables. Using the result of the regression, we derive

predictions regarding the outcome of each game from the 2002 World Cup. When

comparing predictions with results, it turns out that a strategy that consistently picks the

winner of World Cup games using my predictions is more accurate that a strategy that

always picks the favorite implied by the odds offered by the London bookies.

Finally, Chapter 4 determines the probabilities for winning the World Cup for

each team by simulating the predefined structure of games from the World Cup. Also, by

comparing the results of the World Cup simulation to the results of a Round Robin

tournament simulation, the teams with a “lucky draw” are identified.

As far as I know, this is the fist work that draws a parallel between the challenge

of ranking the teams that participate in the soccer World Cup, and the challenge faced

every year by the Bowl Championship Series (BCS) in ranking American college

football. The two tasks are similar in that the number of games played by every team is

relatively small, teams play many more games within their regional league than against

teams in other leagues, and the quality of the opponents varies from region to region.

This thesis modified some of the methods currently used to rank football teams, so that

they rank the soccer teams from the World Cup instead. It is my hope that further

67
 Chapter 5: Conclusion Andrei C. Grecu

research will look at other mathematically rigorous BCS ranking methods and adapt them

to rank the soccer teams that qualify for the soccer World Cup.

Another idea for future research might focus on changing the methods discussed

in this thesis, so that games that take place in the World Cup are immediately

incorporated into the model. Since there is no better indication of the relative value of a

team than its last results, updating the dataset of historical game during the tournament

could make the models even more accurate.

Finally, whenever people are predicting the results of future events, there are

betting implication involved. Since my algorithms were proven to be better predictors of

individual game results than the bookies for the 2002 World Cup games, there might be

“arbitrage opportunities” in the case of the upcoming soccer World Cup games. Of

course, my predictions still have a relatively high rate of failure, so there is no obvious

moneymaking betting strategy. However, here are my predictions for the 2006 World

Cup group games, as well as the simulation results for the probabilities for winning the

World Cup assigned to each team5. Bet at your own risk!

TEAM A TEAM B A WINS B WINS TIE Probability of Winning the Cup

1 Germany Cost Rica 47.75% 24.58% 27.67% Brazil 17.58%

2 Poland Ecuador 22.45% 53.45% 24.10% Italy 11.04%
3 Germany Poland 75.79% 8.19% 16.00% Germany 10.24%
4 Ecuador Costa Rica 25.87% 47.56% 26.58% Portugal 8.42%
5 Ecuador Germany 16.52% 59.87% 23.61% Netherlands 6.96%
6 Costa Rica Poland 64.74% 14.46% 20.80% France 6.32%
7 England Paraguay 37.53% 33.81% 28.66% Argentina 6.20%
8 Trinidad &Tobago Sweden 15.61% 62.75% 21.64% England 4.54%

5
In making the predictions I used the team value from the Eigenvector rankings, which were the most
accurate in 2002.

68
 Chapter 5: Conclusion Andrei C. Grecu

9 England Trinidad & Tobago 66.41% 13.29% 20.29% Sweden 4.16%

10 Sweden Paraguay 33.98% 37.58% 28.44% Paraguay 3.56%
11 Sweden England 32.12% 39.43% 28.45% Switzerland 3.16%
12 Paraguay Trinidad & Tobago 64.56% 14.44% 20.99% Croatia 2.80%
13 Argentina Cote d’Ivoire 68.00% 12.29% 19.70% Spain 2.36%
14 Serbia & Montenegro Netherlands 13.36% 65.77% 20.86% Serbia & Montenegro 2.34%
15 Argentina Serbia & Montenegro 63.54% 14.75% 21.70% Saudi Arabia 1.00%
16 Netherlands Cote d’Ivoire 70.14% 11.06% 18.80% Ecuador 0.98%
17 Netherlands Argentina 37.61% 33.19% 29.20% Ukraine 0.94%
18 Cote d’Ivoire Serbia & Montenegro 11.06% 70.14% 18.80% Mexico 0.80%
19 Mexico Iran 68.88% 11.60% 19.52% Czech Republic 0.48%
20 Angola Portugal 14.54% 64.71% 20.75% Japan 0.38%
21 Mexico Angola 75.60% 8.34% 16.05% USA 0.30%
22 Portugal Iran 57.20% 19.15% 23.65% Iran 0.22%
23 Portugal Mexico 24.82% 47.61% 27.57% Australia 0.20%
24 Iran Angola 44.36% 30.30% 25.34% Angola 0.16%
25 Italy Ghana 73.12% 9.37% 17.50% Ghana 0.10%
26 USA Czech Republic 47.74% 25.90% 26.36% Cote d’Ivoire 0.10%
27 Italy USA 51.23% 21.96% 26.81% Poland 0.08%
28 Czech Republic Ghana 46.83% 27.59% 25.58% Trinidad & Tobago 0.10%
29 Czech Republic Italy 14.41% 63.33% 22.25% Tunisia 0.04%
30 Ghana USA 18.45% 58.36% 23.19% Togo 0.00%
31 Brazil Croatia 55.41% 18.66% 25.92% Korea Republic 0.00%
32 Australia Japan 43.08% 30.06% 26.86% Costa Rica 0.00%
33 Brazil Australia 62.48% 14.48% 23.04%
34 Japan Croatia 24.21% 49.76% 26.02%
35 Japan Brazil 10.91% 69.30% 19.79%
36 Croatia Australia 42.62% 29.80% 27.58%
37 France Switzerland 57.10% 18.69% 24.21%
38 Korea Republic Togo 61.96% 16.37% 21.67%
39 France Korea Republic 46.33% 26.25% 27.42%
40 Togo Switzerland 23.86% 51.85% 24.30%
41 Togo France 10.32% 71.87% 17.80%
42 Switzerland Korea Republic 27.11% 46.47% 26.42%
43 Spain Ukraine 47.53% 25.67% 26.80%
44 Tunisia Saudi Arabia 33.69% 39.25% 27.06%
45 Spain Tunisia 51.55% 22.71% 25.73%
46 Saudi Arabia Ukraine 35.47% 37.21% 27.32%
47 Saudi Arabia Spain 24.95% 48.48% 26.57%
48 Ukraine Tunisia 40.12% 32.83% 27.05%

69
Appendix A: Neural Network C++ Code Andrei C. Grecu

APPENDIX A: NEURAL NETWORK C++ CODE

#include <cstdio>
#include <string>
#include <vector>
#include <iostream>
#include <cmath>
#include <math.h>

#define NMAX 100

#define INFI 0x3f3f3f3f
#define S 50

using namespace std;

const double epsilon = 6.0;

/* prevteam[i] = value of team i at time t-1 */

static double prev[NMAX];

/* currteam[i] = value of team i at time t */

static double curr[NMAX];

/* team names */
vector <string> teams;

/* number of teams */
int N;

int adj[60][60][15];

double update(int i, int j, int k)

{
if (i == j || adj[i][j][k] == 100)
return INFI;

int dif = adj[i][j][k];

if (prev[i] == prev[j])
{
/* win */
if (dif > 0)
{
double maxgain = (0.5-0.005*prev[i]) *prev[j]+
25+0.75*prev[i];
return (maxgain - prev[i])/55.0 * abs(dif*k) + prev[i];
}

/* lose */
if (dif < 0)

70
Appendix A: Neural Network C++ Code Andrei C. Grecu

{
double maxloss = 0.005 * prev[i] * prev[j] + 0.25 *
prev[i];
return (prev[i] - maxloss)/55.0 * abs(dif*k) + maxloss;
}

/* tie */
return prev[i];
}

if (prev[i] > prev[j])

{

/* win */
if (dif > 0)
{
double maxgain = (0.5-0.005*prev[i]) *prev[j]+
25+0.75*prev[i];
return (maxgain - prev[i])/55.0 * abs(dif*k) + prev[i];
}

/* lose */
if (dif < 0)
{
double maxloss = 0.005 * prev[i] * prev[j] + 0.25 *
prev[i];
return (prev[i] - maxloss)/55.0 * abs(dif*k) + maxloss;
}

return 0.005 * prev[i] * prev[j] + 0.5 * prev[i];

/* tie */
double maxloss = 0.005 * prev[i] * prev[j] + 0.25 *
prev[i];
return (2.0*maxloss-prev[i])/12.0*k+prev[i]-
(2.0*maxloss-prev[i])/12.0;
}

if(prev[i] < prev[j])

{
/* win */
if(dif > 0)
{
double maxgain = (0.5-0.005*prev[i]) *prev[j]+
25+0.75*prev[i];
return (maxgain - prev[i])/55.0 * abs(dif*k) + prev[i];
}

/* lose */
if (dif < 0)
{
double maxloss = 0.005 * prev[i] * prev[j] + 0.25 *
prev[i];
return (prev[i] - maxloss)/55.0 * abs(dif*k) + maxloss;
}

71
Appendix A: Neural Network C++ Code Andrei C. Grecu

/* tie */
double maxgain = (0.5-0.005*prev[i]) *prev[j]+
25+0.75*prev[i];
return (maxgain-prev[i])/24.0*k + (25.0*prev[i]-
maxgain)/24.0;
}
}

/* update the team value at time t */

void team_value()
{
for (int i = 0; i < N; ++i)
{
double val = 0.0;
int cnt = 0;

for (int j = 0; j < N; ++j)

for (int k = 1; k <= 13; ++k)
{
double tmp = update(i, j, k);

if (tmp != INFI)
{
val += tmp;
++cnt;
}
}

if (cnt != 0)
curr[i] = (double) val/cnt;
else
curr[i] = prev[i];
}

/* scaling */

double oldmax = -INFI;

double oldmin = INFI;

for (int i = 0; i < N; ++i)

{
oldmax = max(oldmax, curr[i]);
oldmin = min(oldmin, curr[i]);
}

double a = 100.0/(oldmax-oldmin);
double b = - 100.0*oldmin/(oldmax-oldmin);

for (int i = 0; i < N; ++i)

curr[i] = a*curr[i] + b;

72
Appendix A: Neural Network C++ Code Andrei C. Grecu

bool finish()
{
for (int i = 0; i < N; ++i)
if (abs(prev[i] - curr[i]) >= epsilon)
return false;

return true;
}

void stable()
{
int iter = 0;

while (true)
{
if (iter == 1000)
break;

++ iter;

//cerr << ",,,,,,," << endl;

team_value();

if (finish())
break;

/* backup computed values */

for (int i = 0; i < N; ++i)
prev[i] = curr[i];

/* print current iteration */

if(iter<=1){

cout << "Begin Iteration # = " << iter << endl;

for (int i = 0; i < N; ++i)
cout << teams[i] << " " << curr[i] << endl;
cout << "End Iteration" << endl;
}

cout << "# of iterations: = " << iter << endl;

}

int team_to_int(string s)
{
for (int i = 0; i < N; ++i)
if (teams[i].compare(s) == 0)
return i;

73
Appendix A: Neural Network C++ Code Andrei C. Grecu

return -1;
}

void read_data()
{
/* read the # of teams */
cin >> N;

/* read all the teams */

for (int i = 0; i < N; ++i)
{
string s;
cin >> s;

teams.push_back(s);
}

/* read the games */

while (true)
{
string a, b;
int score;
int year;

year = -1;

cin >> a >> b >> score >> year;

if (year == -1)
break;

// cerr << a << " " << b << " " << score << " " << year << endl;

int na = team_to_int(a);
int nb = team_to_int(b);

assert(na != -1);
assert(nb != -1);

year -= 1993;

adj[na][nb][year] = score;
}
}

int main(void)
{
for (int i = 0; i < 60; ++i)
for (int j = 0; j < 60; ++j)
for (int k = 0; k < 15; ++k)
adj[i][j][k] = 100;

74
Appendix A: Neural Network C++ Code Andrei C. Grecu

read_data();

for (int i = 0; i < N; ++i)

prev[i] = S;

stable();

/* print final results */

cout << "FINAL RESULTS" << endl;

for (int i = 0; i < N; ++i)

cout << teams[i] << " " << curr[i] << endl;
cout << "DONE!" << endl;

return 0;
}

75
Appendix B: Neural Network Evolution Andrei C. Grecu

APPENDIX B: NEURAL NETWORK EVOLUTION

After about 50 iterations, the neural network converges to constant values for

every team:

BRA
Neural Network Evolution ITA
FRA
120 ARG
ENG
100
GER
PTG
80
Team Value

NET
60 SPA
SWE
40 CZE
MEX
20 PAR
JAP
0 IRA
0 20 40 60 80 100 IVO
Iterations YUG
POL

76
Bibliography Andrei C. Grecu

BIBLIOGRAPHY

Callaghan, Thomas; Muncha, Peter; and Porter Manson. “Random Walker Ranking for
NCAA Division I-A Football”, Georgia Institute of Technology, June 15, 2005

Callaghan, Thomas; Muncha, Peter; and Porter Manson. “The Bowl Championship
Series: A Mathematical Review”, Notices of the American Mathematical Society,
September, 2004

Cassady, Richard; Maillart, Lisa; Salman, Sinan. “Ranking Sports Teams: A

Customizable Quadratic Assignment Approach”, Interfaces, Vol. 35, No. 6, November-
December 2005

Coleman, Jay. “Minimizing Game Score Violations in College Football Rankings”,

Interfaces, Vol. 35, No. 6, November- December 2005

Colley, Wesley; Colley’s Bias Free College Football Ranking Method, Ph.D. dissertation,
Princeton University

Devore, Jay. Probability and Statistics for Engineering and Statistics, 6th Edition.
Belmont: Brooks/Cole- Thomson learning, 2004.

Dixon, Mark and Coles, Stuart. “Modeling Association Football Scores and Inefficiencies
in the Football Betting Market”, Applied Statistics (1997), 46, No. 2

Dobson, A.J. An Introduction to Generalized Linear Models, Chapman and Hall, London
2001

Dyte D. and Clarke S.R. “A ratings based Poisson model for World Cup soccer
simulation”, Journal of the Operational Research Society (2000) 51, 993-998

Fleiss, Joseph L. Statistical Methods for Rates and Proportions. 2nd Edition, New York:
John Wiley & Sons, 1981.

Goddard, Stephen. “Ranking in Tournaments and Group Decision Making”, Management

Science, Vol. 29, No.12, Dec. 1983

Sharpe, Graham. Gambling on Goals: A Century of Football Betting, Mainstream

Publishing, Edinburgh and London 1997

Jech, Thomas. “The Ranking of Incomplete Tournaments: A Mathematician’s Guide to

Popular Sports”, Department of Mathematics, Pennsylvania State University

Karlis, Dimitris and Ntzoufras, Ioannis. “Statistical Modelling for Soccer Games: The
Greek League”, Department of Statistics, Athens University of Economics and Business

77
Bibliography Andrei C. Grecu

Keener, James. “The Perron-Frobenius Theorem and the Ranking of Football Teams”,
Society for Industrial and Applied Mathematics Review, Vol.35, No.1, pp. 80-93, March
1993

Koning, Ruud. “Balance in Competition in Dutch Soccer”, The Statistician, 49, Part 3, pp
419-431

Martinich, Joseph. “College Football Rankings: Do Computers Know Best?”, College of

Business Administration, University of Missouri- St. Louis, May 2002

McCullagh, Peter and Nelder, J.A. Generalized Linear Models, Chapman and Hall,
London 1989

McGarry T. and Schutz RW. “Efficacy of Traditional Sport Tournament Structures”,

Journal of the Operational Research Society (1997), 48, 65-74

Palomino, Frederic; Rigotti, Luca; Rustichini, Also. “Skill, Strategy, and Passion: an
Empirical Analysis of Soccer”, in progress, Department of Econometrics, Tilburg
University

Pollard R, Benjamin B, and Reep C. “Sports and the Negative Binomial Distribution”,
Optimal Strategies in Sports, 1977

Wilson, Rick. “Ranking College Football Teams: A Neural Network Approach”,

Interfaces 25, July-August 1995 (pp. 44-59)

Winston, Albright. Practical Management Science, Brooks/Cole, 2001

78