Game Intelligence in Team Sports

Abstract

We set up a game theoretic framework to analyze a wide range of situations from team sports. A fundamental idea is the concept of potential; the probability of the offense scoring the next goal minus the probability that the next goal is made by the defense. We develop categorical as well as continuous models, and obtain optimal strategies for both offense and defense. A main result is that the optimal defensive strategy is to minimize the maximum potential of all offensive strategies.

Full text

RESEARCH ARTICLE
Game Intelligence in Team Sports
Jan Lennartsson
1
, Nicklas Lidström
3
, Carl Lindberg
2
*
1Chalmers University of Technology and Gothenburg University, Sweden, 2The Second Swedish National
Pension Fund and Uppsala University, Sweden, 3Independent Researcher, Vasteras, Sweden
*carl@teamsporttactics.com
Abstract
We set up a game theoretic framework to analyze a wide range of situations from team
sports. A fundamental idea is the concept of potential; the probability of the offense scoring
the next goal minus the probability that the next goal is made by the defense. We develop
categorical as well as continuous models, and obtain optimal strategies for both offense
and defense. A main result is that the optimal defensive strategy is to minimize the maxi-
mum potential of all offensive strategies.
Introduction
A subset of all team sports is the ones where two opposing teams each have a goal to defend,
and the team which scores the most points win the game. This paper analyzes general situation
tactics in such sports, henceforth denoted team sports. Game intelligence in team sports is usu-
ally regarded as something very incomprehensible, and excellent players are often praised for
how they “read the game”. Even though most would agree on what constitutes good skills—
technique, strength, agility, endurance, etc—it is less obvious what characterizes a good player
in terms of game intelligence.
We will make an attempt at analyzing the concept of game intelligence from a game theoret-
ic perspective. To this end, we assume that a player’s overall ability can be categorized into two
parts. First, the ability to decide on a strategy, which is in some sense optimal, in each encoun-
tered game situation. Second, to carry out the chosen strategy. The first category is what typi-
cally is contained in the concept of game intelligence, while the last category has to do with a
player’s skill set. The present paper will focus on the first category. Obviously, one can never
know which choice would have worked out the best on each particular occasion. However, we
show in this paper that we can find strategies which are optimal in the mean. To analyze such
situations, we adopt a game theoretic framework. We will model game situations as so called
zero-sum games, i e games where the players have exactly opposite rewards. In our setting, a
goal made by the offense has value one, which is exactly the negative of the value it has for the
defense. Conversely, a goal made by the defense has the value minus one for the offense which
again is exactly the negative of the value it has for the defense. Further, we associate a utility to
each player’s choices. We will define the utility function for team sports to be the potential; the
probability of the attacking team scoring the next goal minus the probability that the next goal
is made by the defending team. Given this setup, we can model a variety of team sport game
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 1/28
a11111
OPEN ACCESS
Citation: Lennartsson J, Lidström N, Lindberg C
(2015) Game Intelligence in Team Sports. PLoS ONE
10(5): e0125453. doi:10.1371/journal.pone.0125453
Academic Editor: MariaPaz Espinosa, University of
the Basque Country, SPAIN
Received: June 12, 2014
Accepted: March 23, 2015
Published: May 13, 2015
Copyright: © 2015 Lennartsson et al. This is an
open access article distributed under the terms of the
Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any
medium, provided the original author and source are
credited.
Data Availability Statement: All relevant data are
within the paper.
Funding: JL was funded by Gothenburg University.
CL was funded by Second swedish Ap fund. The
funders had no role in study design, analysis,
decision to publish, or preparation of the manuscript.
NL received no funding for this work.
Competing Interests: The authors have declared
that no competing interests exist.

situations, and solve these using standard game theoretic methods. An implication of our ap-
proach is that game intelligence is not something incomprehensible, but rather an acquired
skill. Note that our setup is different than the so-called potential games in e g [1]. We have cho-
sen to denote the utility function by potential because of its natural interpretation as a scalar
potential for a conservative vector field.
The first part of the paper gives the theoretical foundation underlying our analysis, includ-
ing the concept of potential fields, and derives some results with applications to game intelli-
gence. The following sections focus on game situations where the players make decisions based
on a given set of strategies. Here, we apply principles from game theory to determine which de-
cisions are optimal. A main consequence of our problem set up is that the optimal defensive
strategy is to make the best offensive choice, in terms of potential, as bad as possible. Further,
the optimal strategy for the offense is to distribute shots between the players so that, for every
player, a shot should be taken if and only if the potential is larger than a certain threshold. This
threshold is the same for all players. It is important to note that the optimal strategy does not
guarantee a successful outcome on each occasion. Rather, the optimal strategy for a specific sit-
uation gives the best outcome in terms of potential.
There is an extensive literature on the application of game theory to fixed game situations in
sports. In [2], the game theoretic analysis of sports as a quantitative field is introduced by test-
ing if professionals use the mixed Nash equilibrium when they decide on which side, forehand
or backhand, to land the first serve. They found that men’s pro tennis players did not random-
ize their serves in an optimal way, but were rather switching the serve direction too often.
Later, [3] found evidence of good strategic play among professionals in the same situation. In
the same setting, [4] included in their analysis also the types of court surfaces, and found evi-
dence that servers on faster surfaces tend to hit the serve to an opponent’s backhand too often.
In [5], a game theoretic approach yields that there is room for improvement both in pitch
selection in Major League Baseball (MLB) as well as in calling plays in the National Football
League (NFL). They can conclude that pitchers throw more fastballs than what is optimal, and
that football teams pass less than they ought to. Interestingly, they find in both sports a nega-
tive serial correlation in the decisions. Further, [2] find the same result in tennis, and [6] report
similar findings for the NFL. This indicates that strategies, in some sports, are changed more
often than what we would see if the strategy choices were truly random. In [6], the authors
offer the explanation that teams excessively switch play types in order to not be perceived as
predictable. The paper [7] present the idea that, in the NFL, the observed temporal dependence
in strategy choice is due to that the offense tries to wear down the defense.
In soccer, [8], [9], [10], and [11] all find that during soccer penalty kicks, both the strikers
and the goalkeepers are making choices that are consistent with the strategies of
Nash equilibria.
Even though fixed game situations play an important role in sports such as ice-hockey, team
handball, basketball and soccer, these sports are primarily built up by in-game activity. Howev-
er, there does not seem to be much literature on quantitative approaches to in-game activity, or
to game intelligence which is the scope of the present paper. The papers [12] and [13] are
slightly related in spirit to the first part of the present paper. In [12], the performance of an bas-
ketball offense is modeled as a network problem. The paper [13] investigates when, in terms of
shot quality, a basketball team should shoot. Further, [14] estimate a matching law from bas-
ketball scoring data, and are able to verify a good fit. Finally, optimal strategies for an underdog
are derived in [15]. In soccer, [16] analyze the in-game decision making of shooting towards
the far post as opposed to the near post.
This paper is organized as follows. In Section Potential, we introduce the concept of poten-
tial, which is of central importance to our analysis. We prove in Section Fundamental results a
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 2/28

theorem and a lemma. These will be used extensively throughout the paper. Sections Strategic
game situations and Extensive game situations present some standard game theoretic notation
and results. Further, we adapt these results to our framework, and give some examples which il-
lustrate the applicability of the theory to game intelligence. In Section Shot potential, we define
the shot potential as the probability of scoring from a given position, and show how this can be
used in the modeling and analysis of various situations in sports. Section Parameter estimation
presents some possible approaches to estimate the model parameters from data. We conclude
with a discussion.
Potential
We introduce in this section the concept of potential. This idea is fundamental to our analysis,
and it is applied in all game situations that we consider.
There are two teams in each game; team A and team B. We define the stochastic expiration
time Tto be the time when the next goal is scored or the game ends. Further, we denote by V
(T) the stochastic variable that takes the value 1 if team A scores, −1 if team B scores, and 0 if
no goal is scored before the game is over.
Definition 1 The potential, v, is defined by
vt¼Et½VðTÞ;8t2½0;T:
Hence, the potential is the probability of team A scoring the next goal, minus the probability
that the next goal is made by team B. The potential vis a general stochastic process in continu-
ous time which is conditioned on the present states of all players, as well as their respective
strategies, skills and tactics. Unfortunately, the potential process vis too complex to model ex-
plicitly. Hence, we will in this paper analyze vonly for a set of game situations which are fre-
quently recurring in their specific sports. Given such a situation, we will be able to find the
optimal player behavior in that particular setting. If the players adopt the optimal behavior,
they improve the potential vfor all t; higher potential from the perspective of team A, and
lower from the point of view of team B. Recall that the potential vconsiders not just the proba-
bility of scoring, but also the effects of losing ball possession.
To illustrate the concept of potential, we consider a few short examples. In a situation where
a team A player has possession with no team B players in her path to the goal, the potential is
close to one. In contrast, if team A has ball possession, but there are a large number of skilled
and well positioned team B players ahead of them, the potential may be close to zero. Further, a
missed pass for team A yields a drop in the potential. This drop could be big for particularly
bad passes, possibly down to almost −1.
We are considering ncategorical alternatives to pursue shot attempts at. Further, the defen-
sive effort,y2[0, 1]
n
, such that ∑y
i
= 1, is the proportional effort that team B puts on each of
the categorical alternatives in order to reduce the potential of the corresponding shot alterna-
tives. In addition, x=(x
1
,...,x
n
)2[0, 1]
n
such that ∑
i
x
i
= 1 is the expected proportion of shot
opportunities from each offensive alternative i=1,...,n. We define the cumulative potential
functions {F
i
}
i=1,...,n
, for each i=1,...,n, as the integral of the corresponding potential fre-
quency function f
i
:[0, 1] × [0, 1] ![−1, 1], such that
Fiðx;yÞ¼Zx
0
fiðx;yÞdx;
for all x,y2[0, 1]. We show now that the potential frequency function can be understood in
terms of the distribution of the quality of the shot opportunities. Consider an offensive alterna-
tive i. We assume that under normal game conditions, the shot opportunities Y
i
(y)2[−1, 1],
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 3/28

given defensive effort yin that shot alternative, are drawn randomly and independently from a
known distribution G
Y
i
(y)
. Here, each value of Y
i
(y) gives the potential for that shot opportuni-
ty. The potential frequency function is defined by
fiðx;yÞ¼G1
YiðyÞð1xÞ;
where ()
−1
denotes the inverse and x,y2[0, 1]. Hence, the potential frequency function gives
the potential in seeking an additional “infinitesimal”proportion of the shot opportunities at
the offensive alternative igiven expected shot proportion xand defense effort y. We see that, by
construction, the potential frequency functions f
i
are monotonically decreasing in the first ar-
gument, respectively. This is natural, since a rational offensive approach is to seek shot at-
tempts at the best shot opportunities first and next consider shot opportunities with smaller
potential. Further, we will assume that the potential frequency functions are also monotonically
decreasing in the second argument, respectively. This models the rational game characteristic
that if more defensive effort is put on shot alternative i, the potential of that alternative is
reduced.
To exemplify, a rational defensive action in response to a high potential categorical offensive
alternative iwould be to increase the defensive effort y
i
, and decrease some or multiple y
j
for j
6¼ i, such that f
i
(x
i
,y
i
) is reduced to the expense of an increase in f
j
(x
j
,y
j
).
We illustrate now the concept of potential with two brief examples, one from ice hockey
and one from team handball.
When is it a good idea to break the rules in a way which causes a 2 minute penalty in ice-
hockey? Assume that the probability that team A scores during a team B penalty is p2(0, 1).
Conversely, the probability that team B scores while being one player short is approximately 0.
Further, we assume that the potential if team A has scored is 0, as is the potential if no team has
scored when the penalty is over. Hence, the potential when the penalty time starts is p. This im-
plies that if it is going to be a good decision for a team B player do to something which gives a 2
minute penalty, her action needs to prevent a situation which had a potential higher than p.
Consider a wing position in team handball, with index w. The wing player shoots when
given possession and a free path up to at least ηmeters up from the short side line. The poten-
tial of shot opportunities when shooting at exactly ηmeters is f
w
(x
w
,y
w
) under the offensive
shot proportions xand defensive efforts y2[0, 1]
n
. If the wing player is supposed to increase
the proportion of expected shots, x
w
, then she needs to take shot opportunities from less than η
meters. Further, the additional shots will have a lower potential, since it is harder to score from
a wide position than from a central one. In addition, if there is an increased defensive effort put
on the wing player, i e if y
w
increases, then she needs to lower the threshold ηin order to main-
tain the same shot proportion x
w
.
Remark 2 Note that it is straightforward to extend the concept of potential to let it account
for various penalties or to sports where a goal can have different value in points depending on
how it was made. Basketball is one example of such a sport.
Fundamental results
Here we state and prove two results, from which a lot of interesting conclusions are drawn.
Consider an offense categorized into nalternatives. We now state the following theorem,
where A
C
denotes the complement of the set A:
Theorem 3 (Tactical Theorem of Team Sports) Suppose that the potential frequency func-
tions {f
i
(x,y)}
i=1,...,n
,for x,y2[0, 1] are continuously differentiable and decreasing in x and y,
respectively, and that @2
@y2fiðx;yÞis non-negative and continuous. Assume that we are given a
fixed defensive effort y 2[0, 1]
n
such that Pn
i¼1yi¼1.Further, the expected shot proportions x
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 4/28

2[0, 1]
n
,subject to Pn
i¼1x
i¼1,satisfies
fjðx
j;yjÞ¼fkðx
k;ykÞ;ð1Þ
for all j,k in some subset K
o
(x)[1, ...,n], with x
i>0if and only if i 2K
o
(x), for
Pj2Kox
j¼1,and where fið0;yiÞ<fjðx
j;yjÞ,for all i 2K
o
(x)
C
and any j 2K
o
(x). Then x
maximizes the potential G defined by
Gðx;yÞ:¼X
n
i¼1
Fiðxi;yiÞ:ð2Þ
Conversely, assume that we are given fixed expected shot proportions x 2[0, 1]
n
,with
Pn
i¼1xi¼1,and a defense y, subject to Pn
i¼1y
i¼1,which satisfies
Zxj
0
@fj
@yj
ðx;y
jÞdx¼Zxk
0
@fk
@yk
ðx;y
kÞdx
for all j, k in some subset K
d
(y)[1, ...,n], with y
i>0if and only if i 2K
d
,where
Pj2Kdy
j¼1.Further,
Zxi
0
@fi
@yi
ðx;0Þdx>Zxj
0
@fj
@yj
ðx;yjÞdx
for all i 2K
d
(y)
C
and j 2K
d
(y). Then yminimizes the potential G(x,y). Finally, if we can
find a point (x,y)for which
Gðx;yÞGðx;yÞGðx;yÞ;
for all feasible x,y,i e a point where xis a maximizer for G when y is fixed at y,and where y
is a minimizer for G when x is fixed at x. Then
max
xmin
yGðx;yÞ¼min
ymax
xGðx;yÞ:
Proof. The first part of the proof follows by noting that a point xwhich satisfies the condi-
tions stated in the theorem allows us to apply the Karuch-Kuhn-Tucker optimization princi-
ples, see e.g. [17, Lemma 14.5]. This gives us that xis a local maximum. Further, Gis concave
with respect to x, since the f
i
are continuously differentiable and decreasing. Hence we have a
concave maximization problem, and it follows from [17, Theorem 2.1], that xis a global
maximizer.
The second part of the proof follows completely analogously.
The final conclusion is an immediate consequence of [17, Lemma 14.8]. ■
Intuitively, it is optimal for the offensive side to always strive to distribute shooting propor-
tions such that all shots which are fired from any position should have a potential larger than
some threshold. This threshold is the same for all offensive alternatives. Hence, if the offense
plays optimally it should be indifferent to which offensive alternative to assign a small addition-
al shot proportion to. On the contrary, the defensive side should seek to distribute its effort
such that if a small defensive effort was added to any defending alternative i, these would yield
the same decrease in F
i
.
We give now a simple but important lemma which helps us to understand how the defense
should optimally position itself. We give the lemma for a single defender, but it is straightfor-
ward to extend it to a multi defender setting.
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 5/28

Lemma 4 (Positioning for Indifference) We assume that the offensive side has nalterna-
tives, each associated with a continuous convex potential function {v
i
(y)}
i=1,...,n
,of the defen-
sive player’sposition, location and velocity, y =(y
l
,y
v
). The offense will pursue its best
alternative, so the potential of the situation v is defined by
vðyÞ¼max
iviðyÞ:
For a set D:={y:g
i
(y)0, 8i=1,...,m}for concave functions g
i
, there exists a global minimizer
y2Dto v.The point yis either a minimum to v
i
for some i=1,...,n, in which case v
j
(y)
v
i
(y)for all j 6¼ i,orysatisfies
vjðy
jÞ¼vkðy
kÞð3Þ
for all j, k in some subset K
d
[1, ...,n], and where v
i
(y)v
j
(y), for all i 2KC
dand any j 2
K
d
.
Proof. The first conclusion follows from [17, Theorem 2.1]. For the second part, if yis the
global minimizer and y6¼ arg min
y2D
v
i
(y) for any i=1,...,n, then for any isuch that v
i
(y)
=v(y) there is a non-zero gradient q
i
such that for small λ>0 we have that y−λq
i
2Dand
v
i
(y−λq
i
)<v(y). Further, since yis the global minimizer and the potential functions are
continuous and convex then there exist a j6¼ isuch that v(y)v(y−λq
i
)=v
j
(y−λq
i
).
Hence, since the potential functions are continuous there exist a j6¼ isuch that v
j
(y)=v(y).
We call an offensive alternative i relevant if v
i
(y)=v(y). If there is only a single element in
the set of relevant alternatives, that option is called the dominating offensive alternative. The
lemma above states that given that there is no dominating offensive alternative, the optimal de-
fense position is such that the potential of at least two, possible several, offensive alternatives
are equal.
Remark 5 Note that in the last parts of the game, the teams may apply a different metric than
the potential, such as simply to maximize the probability of scoring, indifferent of the resulting
potential of the situation. The Positioning for Indifference lemma obviously works regardless of
which underlying functions v
i
we choose, as long as they satisfy the conditions of the lemma.
Notable examples
In this section, we give some examples of applications to the Tactical Theorem of Team Sports
and the Positioning for Indifference lemma. The applications are drawn from ice-hockey and
team handball. However, the results above are not at all constrained to these two sports. Rather,
the two sports were chosen based on the authors personal sporting backgrounds.
Team handball; shot proportions. We will in this example challenge two old “truths”in
team handball. These are that an acceptable level of shot efficiency for wing and pivot players
are about 80%, while an acceptable corresponding level for backcourt players is about 50%. We
have, by the Tactical Theorem of Team Sports, that given a defense y, optimal or not, the opti-
mal expected shot proportions xare such that their potential frequency functions are equal.
Intuitively, the best alternative for each offensive player that they do not pursue all have
equal potential.
Recall that the potential is the probability of team A scoring the next goal, minus the proba-
bility of team B scoring the next goal. Hence, it is different from the probability of team A scor-
ing. In general, the risk of technical faults for reaching a back court or wing shot is less than for
a corresponding pass to the pivot, which is more troublesome to set up.
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 6/28

In order to illustrate the consequence of the Tactical Theorem of Team Sports, we give a
simplified example of a situation with only two offensive players; a back court player and a
wing player. Suppose, given the current defense, that the potential frequency functions for the
back court and wing players are given by f
b
and f
w
, respectively, see Fig 1.
We let x
b
denote the proportion of shots fired by the back court player, and thus x
w
=1−x
b
is the proportion of shots fired by the wing player. Further, we make the reasonable assumption
that the risk of a counter attack from a missed back court attempt is the same as that of a
missed wing attempt. As a consequence, the potential is equal to the shooting efficiency minus
the same constant for both back court and wing players. By the Tactical Theorem of Team
Sports, the optimal proportion of back court player shots is given at the point where the poten-
tial frequency functions satisfy f
b
(x
b
)=f
w
(1 −x
b
). In general, this suggests that the players with
high efficiency should most likely be the ones to pursue additional shot opportunities, at the ex-
pense of the players with lower efficiency. Hence, the more efficient players should be put in
Fig 1. The potential frequency functions f
b
(x
b
) and f
w
(1 −x
b
).
doi:10.1371/journal.pone.0125453.g001
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 7/28

shot positions more often, if that is at all possible without dramatically decreasing their effi-
ciency to levels below those of the less efficient players. The opposite is true for less efficient
players, who should take fewer shots and focus on getting their efficiency up. For a team in op-
timal play, all positions feature equal potential for their best shot opportunity not pursued.
It is natural to assume that the potential frequency functions are convex, meaning in essence
that it is always easier to find bad shot opportunities than good ones. Under this assumption,
the efficiency for all players will be quite close to the potential frequency function threshold at
which to cease taking shots. To illustrate this using the setting of the present example, the effi-
ciency for the back court and wing players will be larger than, but close to, f
b
(x
b
). But this im-
plies that the “truths”stated in the beginning of the example are not truths at all, but rather
signs of a non-optimal team strategy.
The main argument for accepting less efficiency for back court shots is that the defense will
start to move up court if expecting additional shots from those positions. This results in more
space for the efficient pivot players. However, by the Positioning for Indifference lemma, an
optimal defense will only re-position for offensive alternatives that have a potential equal to the
relevant alternatives. Hence, shots that are fired from back court which feature a potential
below some efficiency threshold should not cause the defense to adjust their positions, but rath-
er be appreciated by the defense as a non-optimal offensive strategy.
Team handball; wing change over. Consider a wing change over, which is a frequently oc-
curring offensive opening in team handball. In a wing change over, one of team A’s wings repo-
sitions, with or without ball possession, to the opposite side of the court. The pivot
subsequently screens on the inner side of r1, see Fig 2. We denote the players contributing in
offensive play by: left wing (lw), left side back (lb), center back (cb), right side back (rb), right
wing (rw) and pivot. The defensive positions are numbered from the side: left 1 (l1), left 2 (l2),
left 3 (l3), right 3 (r3), right 2 (r2), right 1 (r1).
The r1 defender chooses either to advance up court, marking the lb player, or to battle the
pivot in an attempt to get around the screening and remove the goalscoring threat from the
screening pivot. In the first case, r2 is responsible for handling the pivot, and these two players
are usually of comparable size and strengths. However, in the latter case, it is the role of r1 to
take care of the pivot. This is typically a mismatch situation where the big pivot player has a
physical advantage over r1. Given r1 stays flat, team A will launch the attack from the lb player.
If r1 moves up court, team A decides between launching the attack from the marked lb player,
or from a central position. We can model this as follows. We denote by v
p
the potential of the
situation where the pivot is screening r1. The potential of an attack started from the lb is v
b
(y),
where yis the position of r1 as the situation begins. Further, v
c
(y) is the potential of the situa-
tion where the attack is started from a central position, without lb. The potential, v
b
(y), de-
creases as r1 approaches lb. However, the farther away r1 is from the remaining part of the
defense, the more sparsely they will have to position themselves to cover the whole 6m line.
Hence, in this case the potential v
c
increases.
In order to illustrate the optimality principle, we let the potential functions be linear in the
distance that r1 lifts from the field line, where v
b
(y) decreases and v
c
(y) increases with respect
to the how far up court r1 moves, see Fig 3.
Given that r1 lifts up court and none of offensive alternatives are dominating, we have by
the Positioning for indifference lemma that the potential of the situation is minimized for y
such that v
b
(y)=v
c
(y). Further, r1 should always lift up court if v
b
(γ)<v
p
, and stay
flat otherwise.
In reality, dichotomous choices to be flat or move up court are typically set at tactics sessions
prior to the game. Here, team tactics are often that the speedy wing defenders should seek to
avoid the screening situation associated with staying flat by choosing the alternative to move
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 8/28

up court. Further, the offensive side attempts to hide the opening, sending over the offensive
wing player lw to its right simultaneously as the lb starts the attack. This has the effect that r1
may not able to reach the optimal position prior to when lb charges. Further, r1 needs to re-cat-
egorize the game situation; to decide to move up court but only to a limited stretch ~
y<y,or
to stay flat and challenge the screening set by the pivot. In the particular game situation dis-
played in Fig 3, then the optimal defensive strategy is to stay flat if ~
yyields a potential
vb~
yðÞ<vp.
Note that the vast majority of wing defenders who choose to advance up court, do this all
the way up to the side back court players, taking them out completely. This will only be the op-
timal defensive strategy given that the corresponding side back court player is the dominating
alternative.
Ice hockey; one against one. In ice hockey, it is a common situation that a single offensive
player with possession faces a single defender. Here, we will draw conclusions from the
Fig 2. Schematic figure of a wing changeover opening in team handball. The offensive players are denoted by triangles and the defensive players are
marked by cirles. The dashed line displays the movement path of lw and the solid lines refer to the possible moves for r1.
doi:10.1371/journal.pone.0125453.g002
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 9/28

Positioning for Indifference lemma in a simplified such game situation. The offensive player,
A
1
, may choose to either shoot or to attempt to dribble past the defender, B
1
. We assume that
regardless of how possession is lost, either by a missed shot or a failed dribble, the resulting po-
tential will be the same. This is a natural assumption, since in either case there are four players
in team A defending the counter attack. An attempt to dribble has a certain probability of fail-
ure, resulting in a loss of possession. However, if it succeeds, A
1
has a free path up to the goal.
This is a situation with a very high probability of scoring, and thus also high potential.
The defending B
1
player has to choose how to position herself in stopping each of the two
offensive alternatives. If B
1
puts much pressure on A
1
early on, by seeking to close the distance
between the players, the offensive choice of shooting will be bad, since A
1
is far away from the
goal and will have to take a shot under pressure. On the other hand, when B
1
puts high pressure
on A
1
, the relative speed between the players will be large. This improves the probability of a
successful dribble, since B
1
will have less time to intercept the puck before A
1
has passed by.
Further, there is much space between the players and the goal for A
1
to use if she chooses this
option. Hence, the dribbling alternative will be better the more pressure B
1
puts on A
1
.In
Fig 3. The potential functions for the example Team handball; wing change over.The potential for a lb start (red) and a center start (blue), respectively,
given that r1 moves up court. For reference, the potential of the situation where r1 stays flat (green) is displayed, even though it is not a function of y.
doi:10.1371/journal.pone.0125453.g003
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 10 / 28

addition, if B
1
chooses to fall back and maintain a low relative speed versus A
1
, the option to
dribble will become worse. Simultaneously, the shot alternative will become better since A
1
can
come closer to the goal undisturbed.
Given that neither of the two offensive alternatives are dominating, by the Positioning for
Indifference lemma B
1
should put pressure on A
1
to the extent that the two options, to shoot or
to dribble, yield the same potential. This is analogous to the red and blue lines in Fig 3. We call
this position y. By playing y,B
1
has minimized the potential of the best choice for A
1
, and
hence plays optimally. Note that A
1
might still score regardless of what B
1
does. However, play-
ing similar situations many times, following the effort ywill result in the least number of goals
scored by team A.
Strategic game situations
In this section, we will set up a strategic game theoretic framework to model various game situ-
ations. Again, there are two teams in the game, team A and team B.
To conduct our analysis we need some definitions and results from game theory, which we
present below. These are completely standard, see e.g. [18]. We are considering strategic zero-
sum game situations, which are of a one-time choice type. Zero-sum games are games for
which the utilities of one side is exactly the negative of that of the other side. We will define the
team A utility function to be the potential, and hence the team B utility function is the
negative potential.
A strategic game situation is a situation where team A and team B each have a choice to
make and none of the participators know the opponent’s actions in advance. Hence, the partic-
ipators make decisions simultaneously and independently, and none of the players have prior
information of their counterpart’s choices.
Definition 6 (Strategic zero sum game) A strategic zero sum game is a structure, hP;G;~
vi,
consisting of the following components:
•two sets of actions, Pand Γ,that team A and team B choose from, respectively.
•a potential function ~
v:PG!½1;1
Further, we associate to a strategic game a set of team A strategies,π2L(P), on the actions
in P, i.e. a team A strategy for a strategic game is a probability distribution on the set of offen-
sive actions. For a finite set P, then π=(π
1
,π
2
,...,π
n
), π
i
0 and ∑
i
π
i
= 1 for the nactions in
P. Analogously, team B has a set of defensive strategies, γ2L(Γ). The potential associated
with the strategies (π,γ) are
vðp;gÞ:¼E½~
vðX;YÞ;
where Xhas distribution π2L(P), and Yhas distribution γ2L(Γ). The notation vdefines the
potential of the game for the pair of strategies (π,γ).
We will have use of the following definitions.
Definition 7 (Nash equilibrium) For a game situation of strategic type, a set of strategies
(π,γ) for team A and team B, respectively, is called a Nash equilibrium if neither side achieves
a higher potential by single-handedly deviating from the strategy, i.e.if
vðp;gÞvðp;gÞ
vðp;gÞvðp;gÞ;ð4Þ
for π2Pand γ2Γ.
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 11 / 28

Definition 8 (max-min strategy) A team A max-min strategy for a strategic game hP;G;~
vi
is the strategy π2Pthat maximizes the function
fðxÞ¼ min
g2LðGÞvðx;gÞ:
The value f(π) is called the team A safety level. Analogously, a team B max-min strategy for the
same game is the strategy γ2Γthat maximizes
gðyÞ¼ min
p2LðPÞfvðp;yÞg;
i e, it maximizes the negative of the potential. Hence,
gðgÞ¼min
g2LðGÞmax
p2LðPÞvðp;yÞ;
so we call γamin-max strategy. The value g(γ) is called the team B safety level.
It can be shown that, at a Nash equilibrium, the team A safety level coincide with the nega-
tive of the team B safety level. We get this and more by the following theorems.
Theorem 9 (Max-Min Theorem)
•For each game situation hP;G;~
vi,there exists a Nash equilibrium.
•A strategy vector (π,γ) in a game situation hP;G;~
viis a Nash equilibrium if and only if π
is a max-min-strategy for team A and γis a min-max-strategy for team B.
•The potential in a Nash equilibrium is equal to the safety level of team A, f(π), and f(π)=−g
(γ).
The Max-Min theorem gives that the concept of Nash equilibria is a stable and satisfying so-
lution to zero-sum games, since both teams can decide their optimal strategy without consider-
ing the strategy of the other team.
The following theorem is a strategic game analogue of the Positioning for
Indifference lemma.
Theorem 10 (Indifference Principle) A game situation with a Nash equilibrium (π,γ)
and potential ~
v has safety level vif and only if
p
i>0)X
j2G
~
vði;jÞg
j¼v
p
i¼0)X
j2G
~
vði;jÞg
jv
g
j>0)X
i2P
~
vði;jÞp
i¼v
g
j¼0)X
i2P
~
vði;jÞp
iv:
Definition 11 (deterministic action) An action a 2L(P) for which a(i)=1for some i =1,
...,n is called a deterministic action.
Definition 12 (strictly dominated action) A team A deterministic action a 2L(P)such
that v(a, γ)<v(π>, γ) for all team B strategies γand some team A strategy π, is called a strictly
dominated action. Conversely, a deterministic action for team B, d 2L(Γ), is strictly dominated
if v(π,d)>v(π,γ) for all team A strategies πand some team B strategy γ.
A consequence of the Indifference Principle is that strictly dominated actions are assigned
probability 0 in a Nash equilibrium.
We have now laid out the theoretic framework for the strategic games in our setting.
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 12 / 28

The Indifference Principle applied to strategic games
We summarize now the results of the previous section in a theorem. The theorem follows di-
rectly from our model setting and standard results of game theory. It states that for a given
game situation, with utility given by the potential, the optimal strategy for both the offense and
the defense is a Nash equilibrium.
Theorem 13 (Fundamental Principle of Strategic Games in Team Sports) For a strategic
game situation hP;G;~
viwith team A strategy π2L(P), and team B strategy γ2L(Γ), the po-
tential is the probability of team A scoring next goal minus the probability of team B scoring
next goal,
vðp;gÞ¼E½VðTÞjp;g:
The Nash equilibrium strategy πfor team A is the strategy that maximizes the minimal po-
tential,
p¼arg max
p2LðPÞmin
g2LðGÞvðp;gÞ:
Further, the Nash equilibrium strategy γfor team B minimizes the maximum potential such
that
g¼arg min
g2LðGÞmax
p2LðPÞvðp;gÞ:
Proof. The result follows immediately from the Max-Min theorem. ■
The strategy γguarantees team B the highest possible potential it can obtain without
knowledge of the team A strategy. By the Max-Min theorem, the Nash equilibrium is given by
the minimax strategy. Hence, analogously to the Positioning for Indifference lemma, the opti-
mal defensive strategy is to make the best alternative for the offense as bad as possible.
Notable examples
In this section we aim to convey the power of applying the Fundamental Principle of Strategic
Games in Team Sports to various game situations in ice hockey.
Here we will break down game situations to categorical strategic games. Recall that a strate-
gic game is a choice situation where the participators make their decisions simultaneously and
have no prior information of the opponents actions.
Ice hockey; chasing the puck. A frequently occurring situation in ice hockey is that a team
A player shoots the puck into her offensive corner, and a player from each team, A
1
and B
1
,
chases after it. At some point, the two players are faced with a choice to either charge forward
in an attempt to win the puck or to hold back and by doing so invite the opposing player to go
first into the situation. This can be modeled as a strategic game defined by
Gc¼hfcharge;hold backg;fcharge;hold backg;~
vi;
where ~
vis the potential defined on each pair of offensive and defensive actions. We denote by
v
A
the potential of the situation where A
1
wins the puck when both players charge. Further, v
B
is the potential when B
1
wins the puck, regardless of the manner by which this happens. Finally,
v
h
is the potential should A
1
charge and B
1
hold back. In this case, B
1
avoids to tackle in the
first instance, and instead positions herself, in balance, within stick reaching distance to A
1
and
the puck. Obviously, v
B
<v
h
<v
A
, since it is better for team A to have puck possession than to
not have it, and since A
1
is in a better position to score if B
1
is not between A
1
and the goal. Ad-
ditionally, pis the probability for A
1
to win the situation given that both players charge and qis
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 13 / 28

the probability for A
1
to win possession given that both players hold back. The game is illustrat-
ed by the matrix given in Table 1.
Since the A
1
action to hold back is strictly dominated by the charge alternative, the Indiffer-
ence Principle yields that the Nash equilibrium contains only the charge action for A
1
. Given
that
vh<pvAþð1pÞvB
then there is a Nash equilibrium at π= (1, 0), γ= (0, 1) and the optimal strategy for A
1
is to
charge while the optimal B
1
strategy is to hold back. Conversely, given that
vh>pvAþð1pÞvB;
there is a Nash equilibrium at π= (1, 0), γ= (1, 0) so it is optimal for both players to charge
in the situation.
In this game situation, intuition leads us to believe that
vBvh<< vA;
since the resulting scoring chance if the offensive player has possession with the defender out
of the way is much higher than for any of the alternative outcomes. Further, if both players
were to choose to charge, they weigh approximately the same, and they come into the situation
with the same speed, then it seems to be approximately a 0.5 probability for B
1
to win the puck
possession. Hence, B
1
should hold back in all such situations.
Ice hockey; pass or dribble. An important choice that all players need to make in many
team sports, is to decide when to pass and when to dribble. We consider this problem for ice
hockey in a strategic game setting. However, the analysis is valid for other sports as well, e g
soccer, team handball, and basketball. Consider a situation where team A has puck possession.
The team A player A
1
with the puck has two choices, to pass or to dribble. Team B on the other
hand can decide between to put pressure on A
1
, or to hold back and wait until team A comes
closer to the team B goal. This typically allows team B to defend themselves in a more compact
and efficient manner, at the expense of that team A can advance forward. The game is defined
by
Gp¼hfpass;dribbleg;fpress;hold backg;~
vi;
where ~
vis the potential defined on each pair of offensive and defensive actions. The matrix for
the game is given in Table 2.
Now, we assume the typical situation that A
1
has sought protection behind her own goal,
and is now advancing with the rest of team A ahead of her. We assume further that v
ph
and v
dh
are such that it is optimal for team B to start to put pressure on A
1
. In addition, we assume that
p>q, which is natural since it is most often considerably easier to succeed with a pass than
with a dribble. Finally, we set v
Bp
>v
Bd
. This implies that the situation resulting from an
Table 1. The matrix of the strategic game G
c
.
team B
charge hold back
team A charge pv
A
+(1−p)v
B
v
h
hold back v
B
qv
h
+(1−q)v
B
doi:10.1371/journal.pone.0125453.t001
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 14 / 28

intercepted pass, v
Bp
, is better for team A than if A
1
is to fail with a dribble, v
Bd
. Obviously, we
have also that v
Ap
>v
Bp
and that v
Ad
>v
Bd
. Given the present situation, v
Ap
v
Ad
, since not
much is won from a successful dribble early on in an attack. If anything, one might suspect that
v
Ap
v
Ad
, as it takes a few seconds to complete a dribble, during which team B has the time to
adjust its positions and put increasing pressure on the rest of team A. This implies that it
would never be optimal to dribble, since pv
Ap
+(1−p)v
Bp
>qv
Ad
+(1−q)v
Bd
. In fact, we need
v
Ap
<<v
Ad
if the decision to dribble is going to be the best choice. Note also that this simple
analysis suggests that given that team B puts pressure on A
1
, she maximizes her potential by
maximizing the probability of a successful pass. Since a successful pass depends not only on the
passer, but also on the receiver and the opponents, the only way to reach a high pass success
rate pis to pass early. The reason is of course that team B will then have little chance to inter-
cept the pass or to force A
1
to dribble.
Extensive game situations
More elaborate sport situations, where the offense and defense make sequential choices, can be
modeled with so-called extensive games. We need to introduce some additional game theoretic
notation for this setting. Again, these are standard, see e g [18].
Definition 14 (Game tree) A game tree hP, p
0
,fiis a structure consisting of the following
components:
•a non-empty set P, the elements of P are called positions;
•an element p
0
2P, the game starting position;
•a function f:P !P(P) from the set of positions to the power set of P i.e. all subsets of P.
The positions in f(p) are called direct followers to p. Furthermore, f has the property that for
each position p 6¼ p
0
there is a unique sequence ðpkÞn
k¼0with p
0
the starting position and p
n
=p
and p
k+1
is a direct follower to p
k
for k = 0, 1, ...,n−1. Any position q which can be reached
from p is called a follower to p. We denote the pair (p, q) consisting of a position and any of its
direct followers a move. The set of ending positions, i.e. positions with no followers, is denoted P
e
and positions in P
i
= P\P
e
are denoted inner positions.
Definition 15 (Extensive game) An extensive game hT;t;fmpgp:tðpÞ¼c;~
viis a structure con-
sisting of the following components:
•a game tree T = hP, p
0
,fi,
•a function t:P
i
!{a, b, c} which determines the turn-order; whether the choice in the position
is to be made is by the team A, team B or a random outcome.
•probability distributions μ
p
over the direct followers of p for all positions such that t(p) = c.
•the potential, ~
v,in the ending positions of the game tree.
Table 2. The matrix of the strategic game G
p
.
team B
press hold back
team A pass pv
Ap
+(1−p)v
Bp
v
ph
dribble qv
Ad
+(1−q)v
Bd
v
dh
doi:10.1371/journal.pone.0125453.t002
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 15 / 28

Extensive games may be illustrated in game trees where the turn order is displayed. Here, di-
rect followers are displayed by connections between the nodes, and triangles denote actions in
continuum (such as choice of position or velocity). Further, lines denote categorical actions,
such as to shoot or to pass, see Fig 4.
Definition 16 (Subgame) For p 2P, the restriction of a game G starting in p is called a sub-
game G
p
to G.
Definition 17 (Team positions) The positions p 2P
i
such that t(p) = a are called team A po-
sitions and denoted P, analogously are Γ={p2P
i
:t(p) = b} called team B positions, and Λ={p
2P
i
:t(p) = c} called chance positions.
Definition 18 (Strategy) Ateam A strategy,σ
A
:P7! L(P)in a game situation on extensive
form is a function with the property that σ
A
(p) assigns probabilities to all direct followers q 2f
(p), for every team A position p. I.e. σ
A
gives the probabilistic decision path for team A, for all its
possible positions. Analogously, a team B strategy,σ
B
, defines the corresponding probabilistic de-
cision path of team B.
We will denote by L(P) and L(Γ) the set of all team A and team B strategies, respectively.
The potential of the game, given strategies πand γ, is set to be
vðp;gÞ¼E½~
vðX;YÞ;
where Xhas distribution π2L(P) and Yhas distribution γ2L(Γ).
Fig 4. A game tree. An extensive game where p
0
is the starting position, and p
0
(team B), p
1
(team A) and p
2
(chance) are inner positions. Here, team B has a decision in continuum while the team A and chance moves
are categorical.
doi:10.1371/journal.pone.0125453.g004
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 16 / 28

Definition 19 (Nash equilibrium for extensive game situation) For an extensive game
hT;t;fmpgp:tðpÞ¼c;~
vi,strategy vectors (π,γ) such that π=L(P)and γ=L(Γ)are called a
Nash equilibrium if
vðp;gÞvðp;gÞvðp;gÞ;
for all strategies π=L(P)and γ=L(Γ). Further, if the restriction of π,γto every subgame is
a Nash equilibrium in that game, it is called a subgame perfect Nash equilibrium.
The Indifference Principle applied to extensive games
We summarize here the results of the previous section in a theorem, which is an immediate
consequence of our game theoretic model. The theorem states that for a given game situation,
with utility given by the potential, the optimal strategy for both the offense and the defense is a
subgame perfect Nash equilibrium.
Theorem 20 (Fundamental Principle of Extensive Games in Team Sports) For an exten-
sive game situation G ¼hT;t;fmpgp:tðpÞ¼c;~
vithe subgame perfect Nash equilibrium strategy π
for team A is the strategy that maximizes the minimal potential,
p¼arg max
p2LðPÞmin
g2LðGÞvðp;gÞ;
for every subgame G
p
. Further, the subgame perfect Nash equilibrium strategy γfor team B
minimizes the maximal potential such that
g¼arg min
g2LðGÞmax
p2LðPÞvðp;gÞ;
for every subgame G
p
.
Proof. The result follows immediately from the definition of subgame perfect
Nash equilibrium. ■
Completely analogously to the strategic game setting, the strategy γguarantees that team B
gets the lowest possible potential it can achieve without knowing anything about the team A
strategy. Hence, the result above extends the Positioning for Indifference lemma such that for
sequential moves, the optimal defensive strategy is to make the best alternative for the offense
as bad as possible.
Notable examples
We analyze now a few examples where each team can make sequential choices.
Ice hockey; two against one. In ice hockey, the game situation where two team A players
A
1
and A
2
face a single team B defender B
1
occurs often. We assume that A
1
has initial puck
possession. Due to the blue line offside, the situation starts as a one against one situation where
B
1
positions herself with the move (p
0
,p
1
) to prevent the attack. Next, A
1
can choose either of
the following four moves: to dribble (p
1
,p
2
); to shoot (p
1
,p
3
); to avoid B
1
,(p
1
,p
4
); and to pass
A
2
,(p
1
,p
5
). If A
1
decides to dribble then she is either successful, (p
2
,p
6
), or looses possession,
(p
2
,p
7
). Further, if A
1
shoots, she ends up in p
3
, an ending position. If A
1
makes the move to
avoid B
1
,(p
1
,p
4
), then B
1
re-positions with the move (p
4
,p
8
). The subgame situation is ended
by a shot from A
1
,(p
8
,p
9
), or with a pass to A
2
for her to shoot a one timer, (p
8
,p
10
). Finally, if
A
1
passes A
2
,(p
1
,p
5
), we are in a new two against one situation, but where A
2
has puck posses-
sion. The corresponding game is illustrated in Fig 5, where a description of the positions are
given in Table 3.
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 17 / 28

Fig 5. The extensive game tree in the example Ice hockey; two against one.
doi:10.1371/journal.pone.0125453.g005
Table 3. Description of positions for the example Ice hockey; two against one.
Positions Description
p
0
team B position, chooses a position in continuum.
p
1
team A position, chooses categorically to dribble, to shoot, to pass, or to take a new position to
avoid confrontation.
p
2
chance position, dribble move.
p
3
ending position, shot by A
1
.
p
4
team B position, chooses a position in continuum.
p
5
pass to A
2
, the game starts over in p
0
with newpositions.
p
6
ending position, successful dribble.
p
7
ending position, failed dribble.
p
8
team A position, chooses categorically to shot or to pass.
p
9
ending position, shot by A
1
.
p
10
ending position, shot by A
2
.
doi:10.1371/journal.pone.0125453.t003
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 18 / 28

Note that there is a certain amount of subjectivity regarding how many nodes to include,
and which ones.
As in the one against one example, we assume that regardless of how possession is lost, the
resulting potential will be the same. Note that if the potential of the subgames starting in posi-
tions p
4
and p
5
are dominated by the subgames started in the alternatives p
4
and p
5
, the present
game will be played identically to the one against one example. However, in reality the opposite
is true; A
1
knows that team A benefits considerably from being two against one, rather than
one against one, so a rational A
1
will avoid the moves (p
1
,p
2
) and (p
1
,p
3
) in the first instance.
The player B
1
knows this, too, and can therefore put more pressure on A
1
in order to force A
1
towards the side of the rink, to a position with smaller potential. The defender B
1
can do this to
the extent that a pass back to A
2
will yield an equally good potential. Hence, to pass and to
avoid are the relevant alternatives for A
1
in node p
1
. The potential of the ending situation is
strongly dependent on the position of A
1
in node p
4
. The subgame started at p
4
will have
smaller potential the further out to the side A
1
has been forced by B
1
. Next, due to the limited
mobility of the goalkeeper, the position p
10
, a pass to A
2
, will have a very high potential. Con-
versely, the goalkeeper can save a large proportion of the shots coming from A
1
, due to that she
has clear vision and can attain good positioning. Hence, since it is easier for B
1
to intercept a
pass if she is near A
2
, the optimal position for B
1
will be close to A
2
, focusing primarily on pre-
venting the pass from A
1
to be completed. This is done to the extent that the potential of the
two alternatives are equal, by the Positioning for Indifference lemma. Consequently, the de-
fender B
1
should always maintain focused on A
2
until A
1
is close enough to the goal so that B
1
can put pressure on both A
1
and A
2
simultaneously. It follows that the optimal team B defender
trajectory will be shaped like an S, see Fig 6.
The player B
1
starts by putting aggressive pressure on A
1
—more so than if A
1
was to come
alone in a one against one in the same position—to force her to the side. She then withdraws to
devote her main attention to prevent a pass to A
2
. Finally, B
1
comes back to a position in front
of her goal keeper. From here, B
1
can both intercept a pass to A
2
and stop A
1
from advancing
closer to the goal with the puck at the same time.
Rationalizable beliefs
We indicate now a direction in which it is natural to expand the framework that we have devel-
oped so far.
We know that players deviating from the Nash equilibrium will invite the counterpart to
improve the potential. For example, if team B knows that team A will follow a certain strategy
^
p2LðPÞ, then team B can lower the potential of the game situation by choosing the optimal
defense arg ming2LðGÞvð^
p;gÞ. Hence, it is distinctly good to have the ability to choose “late”in
decision situations, which is a skill that professional players practice in many sports.
To illustrate further, consider the ice hockey example of one against one. If the defender
knows that the offensive player is reluctant to dribble, then she can put more pressure early on,
in order to stear the offensive player further away and to the sides. This renders that the poten-
tial of the game will be smaller than if team A played by the optimal Max-Min strategy.
As an example from team handball, we re-visit the “truth”that the wing players should have
a much higher efficiency than the back court players. We argued earlier that this is likely to be
a suboptimal strategy. If we assume that team A takes too few shots from some positions, then
team B can focus more effort on making the remaining positions even worse, which has the ef-
fect that the players who take too large shot proportions will have trouble to maintain their
shot efficiency.
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 19 / 28

Shot potential
We introduce in this section the concept of a shot potential, which is related to the potential.
The shot potential fields can be used to make numerical analysis of various game situations.
The present paper is to a large extent based on the concept of potential i.e. the probability of
team A scoring next minus the probability of team B scoring next. However, in many situations
the potential will be approximately equal to the probability of team A scoring, see e.g. the 2
against 1 example in Section 1. Thus the potential depends only on the probability of scoring,
given the chosen strategies for each side. We make the following definitions.
Fig 6. Outline of the optimal trajectories in the example Ice hockey; two against one.The offensive players are denoted by triangles, where a solid
triangle marks the possession holder, and the defender is denoted by a circle. In the upperrectangle, B
1
puts pressure on the puck holder A
1
, who avoids
confrontation to await A
2
. In the lower rectangle, having steared A
1
into a worse position, B
1
shifts focus to A
2
, in order to decrease the greater threat posed
by that player.
doi:10.1371/journal.pone.0125453.g006
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 20 / 28

Definition 21 (Shot potential) The shot potential for A
1
is the probability that A
1
scores
with an instant shot from her present position, given that the resulting situation if the shot is
missed has potential 0.
Definition 22 We refer to the level curves of the shot potential as isolines.
Note that the isolines give sets of points for which the shot potential is equal.
Notable examples
Here we will use the shot potential to analyze several situations in ice hockey in a
dynamic setting.
Ice hockey; eccentric isolines. Consider a one against one situation in ice hockey. We as-
sume that the team A attacker A
1
is skating towards the goal with speed v
A
. The attacker A
1
can not shoot if she is closer than rto the team B defender B
1
, since B
1
can then interfere with
the shot. Further, if the attacker A
1
decides to try to skate past B
1
outside her reach, and the iso-
line which she is presently skating along is circular with radius r
A
, what is the necessary speed
v
B
of B
1
that can guarantee that A
1
will not be able to obtain a higher shot potential before
reaching the goal? We see immediately that v
B
must satisfy
vA
vB
rA
rAr;
so that vBvArArðÞ
rA.IfB
1
can keep this critical speed, then she should attempt to hold A
1
at the
present isoline. If not, B
1
should fall back and hold an isoline associated with a larger shot po-
tential, see Fig 7.
Remark 23 Note that the team B strategy in the example above minimizes the maximal shot
potential during the course of the situation.
Ice hockey; parametric isolines. Beside the skills of the shooter, the probability of scoring
with a shot in any team sport depends on at least two factors; the distance to the goal and the
firing angle. Obviously, the chance of scoring will improve the closer to the goal, and the more
central the position, the shot is fired from.
We derive here two simple parametric models of the shot potential. The first one is for a
team A puck holder. The second is for another team A player who does not have puck posses-
sion and who shoots instantly when she get the puck—a so called one timer. We define the
shot potential model for the puck holder as
fðx1Þ¼aelr1cos ðy1Þ;
and the shot potential for the player who shoots a one timer as
gðx1;x2Þ¼ðaþbjy1y2jÞelr2cos ðy2Þ:
Here x
i
,i= 1, 2, are the locations for A
i
,(r
i
,θ
i
) is the polar representation of x
i
around the y-
axis for a coordinate system with origo in the middle of the goal, λ,α,β>0, and α+βπ1.
The players’shot potentials are illustrated in Figs 8and 9.
The motivation for the second model is that a one timer may have a considerably larger
shot potential than if that shot would have been fired from a player who has had puck posses-
sion for some time. The reason is that, in the one timer case, the goalkeeper needs to make a
sudden shift of position, while for a direct shot, the goalkeeper is already well positioned to
save the puck as the shot is fired.
Ice hockey; optimal trajectories. Here we use the parametric shot potential model defined
in Section Ice hockey; parametric isolines to derive the optimal trajectories for a simple example
of a two player offense facing a single defender. All trajectories are given on a discretized grid.
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 21 / 28

Recall that the game tree of this setting is given in Fig 5. We assume that the A
1
move to drib-
ble, (p
1
,p
2
), is dominated by the other alternatives. The player B
1
makes the first move.
The positions for the participating players are denoted by pði;jÞ
X, where Xdenotes the player, i
is the level starting from the top, and jgives the node in that level counted from the left. We as-
sume that B
1
is the center point of a ball with radius rand that the offensive players keep at
least that distance to B
1
at all nodes. If B
1
is close to the forward who is receiving a pass, then B
1
intercepts that pass with probability q
c
2[0, 1]. Otherwise, B
1
intercepts passes with probability
q
p
2[0, 1], where q
c
>q
p
. The potential of a shot fired by a puck possession holder Xis given
by fðpi;jðÞ
XÞ. Similarly, the shot potential of a one timer from Yfollowing a pass from player Xis
gp
i;jðÞ
X;pk;lðÞ
Y

.
The time dynamic locations for the players may be illustrated in game trees. These trees de-
scribe the trajectories for each player, and depend on the moves of B
1
. The direct followers to
Fig 7. The optimal B
1
strategy for the example Ice hockey; eccentric isolines.The offensive starting position is denoted by a triangle and the defensive
positions are denoted by circles. Here B
1
has critical speed vB¼vAðrArÞ
rAand the present isoline is circular.
doi:10.1371/journal.pone.0125453.g007
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 22 / 28

the inner positions of the game tree for each player are listed in Table 4. The ending positions
are given in Table 5.
Note that since it is easier for B
1
to intercept a pass if she is close to the pass receiving for-
ward, the optimal ending position for B
1
must always be such that she puts pressure on both of-
fensive players simultaneously. For this reason, this ending position is the only one we will
include for B
1
on the final level.
The grid for a specific example is given in Fig 10, where the filled circles and triangles indi-
cate the optimal strategies. The trajectory of A
1
depends on how much pressure B
1
puts at level
1. Similarly, the trajectory of A
2
is determined by the position of B
1
at level 2. Note that the op-
timal strategy is S shaped, which is in line with the reasoning in Section Ice hockey; two against
one.
Remark 24 We have chosen a sparse grid for simplicity of exposition. However, it is straight-
forward to consider more dense ones. This would allow for more realistic modeling. Further, our
results in the example are robust to variations in how the nodes are positioned.
Fig 8. The isolines of the shot potential of A
1
in the example Ice hockey; parametric isolines.The parameter values λ= 0.03, α= 0.2.
doi:10.1371/journal.pone.0125453.g008
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 23 / 28

Parameter estimation
We have in this paper derived game theoretic models which are general enough to cover a wide
range of game situations in several team sports. The applicability of the theory will depend,
among other things, on how well we can statistically estimate the model parameters. To this
end, we will in this section briefly describe some possible approaches to parameter estimation.
Statistical analysis of game data
We have in previous sections analyzed a number of game situations which are frequently oc-
curring in their respective sport. By suitably categorizing entire games into a number of such
situations, what choices the players made, and the outcome, it is in principle straightforward to
obtain estimates of the model parameters for those particular situations. By doing this, one can
Fig 9. The isolines of the shot potential of A
2
, given that the pass comes from A
1
in the example Ice hockey; parametric isolines.The assisting player
A
1
is positioned at the green triangle and the parameter values are λ= 0.03, α=β= 0.2.
doi:10.1371/journal.pone.0125453.g009
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 24 / 28

conclude whether teams as a whole, or even specific players, appear to play in a non-minimax
optimal way.
One against one. Previously in this paper, we have addressed this game situation in an ice
hockey setting. However, it is very common in many other sports as well, including basketball,
soccer, and team handball. Given a specific sport, we assume that we have observed a number
of games for team B. In these games, we have noccasions which we have judged to be one
against one situations. Further, we categorize the choices that the offensive players did in each
situation into either shoot or dribble. We have now n
s
data points where the offensive player
chose to shoot, and n
d
situations where she decided to dribble. The number of goals scored
Table 4. The direct followers to inner positions, with location given by pði;jÞ
X, for a player X.
Inner positions Direct followers
pð0;1Þ
B1pð1;1Þ
B1;pð1;2Þ
B1;pð1;3Þ
B1;pð1;4Þ
B1on
pð1;1Þ
B1pð2;1Þ
B1;pð2;2Þ
B1;pð2;3Þ
B1on
pð1;2Þ
B1pð2;2Þ
B1;pð2;3Þ
B1;pð2;4Þ
B1;pð2;5Þ
B1on
pð1;3Þ
B1pð2;3Þ
B1;pð2;4Þ
B1;pð2;5Þ
B1;pð2;6Þ
B1on
pð1;4Þ
B1pð2;4Þ
B1;pð2;5Þ
B1;pð2;6Þ
B1on
pð2;4Þ
B1;pð2;5Þ
B1;pð2;6Þ
B1pð3;1Þ
B1on
pð0;1Þ
A2jpð1;jÞ
B1on pð1;jÞ
A1on
pði;jÞ
A1pðiþ1;jÞ
A1o;i>0
n
pð0;1Þ
A2pð1;1Þ
A2on
pð1;1Þ
A2jnpð2;1Þ
B1;pð2;2Þ
B1;pð2;3Þ
B1;pð2;4Þ
B1opð2;1Þ
A2;pð2;2Þ
A2;pð2;3Þ
A2on
pð1;1Þ
A2jpð2;5Þ
B1on pð2;2Þ
A2;pð2;3Þ
A2on
pð1;1Þ
A2jpð2;6Þ
B1on pð2;3Þ
A2on
pð2;1Þ
A2jpð2;1Þ
B1;pð2;2Þ
B1;pð2;3Þ
B1on pð3;1Þ
A2;pð3;2Þ
A2;pð3;3Þ
A2;pð3;4Þ
A2on
pð2;1Þ
A2jpð2;4Þ
B1on pð3;2Þ
A2;pð3;3Þ
A2;pð3;4Þ
A2on
pð2;2Þ
A2pð3;3Þ
A2;pð3;4Þ
A2on
pð2;3Þ
A2pð3;4Þ
A2on
doi:10.1371/journal.pone.0125453.t004
Table 5. The ending positions for each player’s game tree.
Player Ending positions
B
1
pð2;1Þ
B1;pð2;2Þ
B1;pð2;3Þ
B1;pð3;1Þ
B1
A
1
pð3;1Þ
A1;pð3;2Þ
A1;pð3;3Þ
A1;pð3;4Þ
A1
A
2
pð3;1Þ
A2;pð3;2Þ
A2;pð3;3Þ
A2;pð3;4Þ
A2
doi:10.1371/journal.pone.0125453.t005
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 25 / 28

when the offensive players decided to shoot and dribble is denoted by X
s
and X
d
, respectively.
Hence, X
s
*Bin(n
s
,p
s
) and X
d
*Bin(n
d
,p
d
). We can now apply standard statistics to draw
conclusions. E g, we can investigate whether we can reject the hypothesis that the probability of
scoring is the same for the shoot and the dribble alternative, respectively. If we are able to reject
this hypothesis, we can also conclude that team Bdoes not appear to play the optimal min-
max strategy.
Experimental design
Given a game situation, it is reasonable to aim to find the overall optimal strategy. However,
due to the complexity of many game situations, it is likely to be insufficient to merely analyze
game data. The reason is that this approach only has in its scope to discover if an existing strat-
egy is better than some other existing strategy. It will be less efficient at determining what is
Fig 10. Example of a grid and the corresponding optimal trajectories for A
1
,A
2
, and B
1
, in the example Ice hockey; two against one.The offensive
players are denoted by triangles, where a solid triangle marks the possession holder, and the defender is denoted by a circle. We use the shot potential
model in Section Ice hockey; parametric isolines with parameters α=β= 0.2, λ= 0.03, q
p
= 0.1, and q
c
= 0.5.
doi:10.1371/journal.pone.0125453.g010
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 26 / 28

indeed the actual optimal strategy. To be able to obtain this, we need to consider
experimental design.
Parameterizing the isolines. While we, thus far, have discussed isolines mainly in relation
to ice hockey, the concept is central to e g basketball, soccer and team handball as well. The rea-
son is that the isolines are important factors in determining how to defend in one against one,
two against one and many other common game situations. To estimate the isolines in a con-
trolled experiment is particularly straightforward. Random players take shots from random po-
sitions, in relevant cases at an undisturbed goalkeeper. The outcome (goal or no goal), as well
as the position from which the shot was taken, is recorded. To estimate the isolines can now be
done by e g applying standard logistic regression techniques.
Ice hockey; chasing the puck. It is straightforward to set up a controlled experiment for
the example Ice hockey; chasing the puck. In the setting of the experiment, we can for each
team Bdefender B
1
run the game situation multiple times. For each situation, B
1
gets assigned
if she should either play the alternative to charge or if she should use the other option; to hold
back. The offensive player A
1
, who is assigned at random for each situation, will always charge,
by the analysis in the example Ice hockey; chasing the puck. We can run the experiment a suffi-
ciently large number of times to be able to conclude to what extent either of the defensive alter-
natives are better than the other. Examples similar to the present one can be found in several
other team sports.
Discussion
In this paper, we make an attempt to develop a mathematical theory on game intelligence in
team sports. It is central to this theory to value game situations by their potential. Intuitively,
the potential is the probability that the offense scores the next goal minus the probability that
the next goal is made by the defense. We give many examples to illustrate the width of the ap-
plicability of our results, but the set of chosen situations is by no means exhaustive.
In Section Team handball; shot proportions, we argue that the classical efficiency thresholds
—e g that back court and wing players should have 50% and 80% mean shot efficiency, respec-
tively—are likely to be non-optimal. It would be interesting to investigate this issue further.
One of the authors, Nicklas Lidström, relied on a set of first principles which he used to ana-
lyze how to play game situations during his career as a professional ice hockey player. His ap-
proach constitutes a cornerstone in the present paper.
In Section Ice hockey; chasing the puck, we analyzed whether a defense player should charge
or hold back. We note here that it appears that the vast majority of backchecking hockey play-
ers judge that it is optimal to charge. Nicklas thought that it was optimal for him to hold back,
which consequently was how he played in such situations.
The example in Section Ice hockey; pass or dribble presents Nicklas’analysis of why it is opti-
mal to pass early in that situation, and in similar ones.
Further, in one against one situations, Nicklas recalls using his reached out stick extensively
as a first line of defense against opposing forwards. He believed that this increased the width
with which he could operate. The operating width is equivalent to the quantity rin the exam-
ples of this paper. Hence he made the action to dribble less attractive in terms of potential for
the opposing forwards. This had the effect that he could stear the forward further to the sides—
preferably their back hand side—to make that alternative, too, lower in potential than what he
thought he could have obtained otherwise.
We note that Nicklas’S shaped strategy in the two against one examples in ice hockey is dif-
ferent to how most defense players choose to play such situations. It seems that most defenders
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 27 / 28

eventually decide to let go of the non-puck holding forward to focus on the forward with puck
possession instead.
Acknowledgments
The authors are grateful to the academic editors and a reviewer for insightful comments which
helped improve the paper considerably. The third author would also like to thank Anders Lor-
demyr for fruitful input concerning the applications of our results to soccer.
Author Contributions
Wrote the paper: CL JL NL. Mathematical modeling and analysis: CL JL NL.
References
1. Monderer D, Shapley LS (1996) Potential games. Games and Economic behavior 14(1): 124–143. doi:
10.1006/game.1996.0044
2. Walker M, Wooders J (2001) Minimax play at Wimbledon. American Economic Review 91(5): 1521–
1538. doi: 10.1257/aer.91.5.1521
3. Hsu SH, Huang CY, Yang CT (2007) Minimax play at Wimbledon: Comment. American Economic Re-
view 97(1): 517–523. doi: 10.1257/aer.97.1.517
4. Bailey BJ, McGarrity JP (2012) The effect of pressure on mixed-strategy play in tennis: The effect of
court surface on service decisions. International Journal of Business and Social Science 3(20): 11–18.
5. Kovash K, Levitt S (2009) Professionals do not play minimax: Evidence from Major League Baseball
and the National Football League. NBER Working Papers No. 15347.
6. Emara N, Owens D, Smith J, Wilmer L (2014) Minimax on the gridiron: Serial correlation and its effects
on outcomes in the National Football League. MPRA Paper No 58907.
7. Mitchell M (2010) Dynamic matching pennies with asymmetries: An application to NFL play-calling.
Working Paper.
8. Chiappori PA, Levitt S, Groseclose T (2002) Testing mixed-strategy equilibria when players are hetero-
geneous: The case of penalty kicks in soccer. American Economic Review 92(4): 1138–1151. doi: 10.
1257/00028280260344678
9. Palacios-Huerta I (2003) Professionals play minimax. Review of Economic Studies 70: 395–415. doi:
10.1111/1467-937X.00249
10. Buzzacchi L, Pedrini S (2014) Does player specialization predict player actions? Evidence from penalty
kicks at FIFA world cup and UEFA euro cup. Applied Economics 46(10): 1067–1080. doi: 10.1080/
00036846.2013.866205
11. Azar OH, Bar-Eli M (2011) Do soccer players play the mixed-strategy Nash equilibrium? Applied Eco-
nomics 43(25): 3591–3601. doi: 10.1080/00036841003670747
12. Skinner B (2010) The price of anarchy in basketball. Journal of Quantitative Analysis in Sports 6(1):
595–608. doi: 10.2202/1559-0410.1217
13. Skinner B (2012) The problem of shot selection in basketball. PLoS ONE 7(1) e30776. doi: 10.1371/
journal.pone.0030776 PMID: 22295109
14. Alferink LA, Critchfield TS, Hitt JL, Higgins WJ (2009) Generality of the matching law as a descriptor of
shot selection in basketball. Journal of Applied Behavior Analysis 42(3): 595–608. doi: 10.1901/jaba.
2009.42-595 PMID: 20190921
15. Skinner B (2011) Scoring strategies for the underdog: Using risk as an ally in determining optimal sports
strategies. Journal of Quantitative Analysis in Sports 7(4) 11. doi: 10.2202/1559-0410.1364
16. Moschini G (2004) Nash equilibrium in strictly competitive games: live play in soccer. Economics Let-
ters 85: 365–371. doi: 10.1016/j.econlet.2004.06.003
17. Nash SG, Sofer A (1996) Linear and nonlinear programming. The MacGraw-Hill Companies, Inc.
18. Osborne MJ (2009) An Introduction to Game Theory. Oxford University Press, international edition.
Game Intelligence in Team Sports
PLOS ONE | DOI:10.1371/journal.pone.0125453 May 13, 2015 28 / 28