Submitted to the Annals of Applied Probability

CONNECTIVITY AND EQUILIBRIUM IN RANDOM GAMES

BY CONSTANTINOS DASKALAKIS

,
ALEXANDROS G. DIMAKIS

AND ELCHANAN MOSSEL

We study how the structure of the interaction graph of a game affects the
existence of pure Nash equilibria. In particular, for a xed interaction graph,
we are interested in whether there are pure Nash equilibria arising when ran-
dom utility tables are assigned to the players. We provide conditions for the
structure of the graph under which equilibria are likely to exist and comple-
mentary conditions which make the existence of equilibria highly unlikely.
Our results have immediate implications for many deterministic graphs and
generalize known results for random games on the complete graph. In partic-
ular, our results imply that the probability that bounded degree graphs have
pure Nash equilibria is exponentially small in the size of the graph and yield a
simple algorithm that nds small non-existence certicates for a large family
of graphs. Then we show that in any strongly connected graph of n vertices
with expansion (1 +(1)) log
2
(n) the distribution of the number of equilib-
ria approaches the Poisson distribution with parameter 1, asymptotically as
n +.
In order to obtain a rened characterization of the degree of connectiv-
ity associated with the existence of equilibria, we also study the model in
the random graph setting. In particular, we look at the case where the in-
teraction graph is drawn from the Erd os-R enyi, G(n, p), model where each
edge is present independently with probability p. For this model we estab-
lish a double phase transition for the existence of pure Nash equilibria as
a function of the average degree pn, consistent with the non-monotone be-
havior of the model. We show that when the average degree satises np >
(2+(1)) log
e
(n), the number of pure Nash equilibria follows a Poisson dis-
tribution with parameter 1, asymptotically as n . When 1/n << np <
(0.5(1)) log
e
(n), pure Nash equilibria fail to exist with high probability.
Finally, when np = O(1/n) a pure Nash equilibrium exists with constant
probability.

EECS, MIT. Supported by a Sloan Fellowship in Computer Science, and NSF award CCF-
0953960 (CAREER). Part of this work was done while the author was at EECS, UC Berkeley, with
the support of a Microsoft Research Fellowship, and NSF awards CCF-0635319, DMS 0528488 and
DMS 0548249 (CAREER).

EECS, USC. Part of this work was done while the author was at EECS, UC Berkeley. Supported
by NSF award DMS 0528488 and a Microsoft Research Fellowship.

UC Berkeley and the Weizmann Institute of Science. Supported by a Sloan fellowship in Math-
ematics, NSF awards DMS 0528488 and DMS 0548249 (CAREER), and ONR grant N0014-07-1-
05-06.
AMS 2000 subject classications: Primary 91A; secondary 60, 68Q
Keywords and phrases: Game Theory, Graphical Games, Connectivity, Phase Transitions, Ran-
dom Constraint Satisfaction Problems
1
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
2
1. Introduction. In recent years there has been a convergence of ideas from
computer science and the social sciences aiming to model and analyze large com-
plex networks such as the web graph, social networks and recommendation sys-
tems. From the computational perspective, it has been recognized that the suc-
cessful design of algorithms performed on such networks, including routing, rank-
ing and recommendation algorithms, must take into account the social dynamics
and economic incentives as well as the technical properties that govern these net-
works [20, 24, 27].
Game theory has been very successful in modeling strategic behavior in large
systems of economically incentivized entities. In the context of routing, for in-
stance, it has been employed to study the effect of selshness on the efciency of
a network, whereby the performance of the network at equilibrium is compared
to the performance when a central authority can simply dictate a solution [7, 30
32]. The effect of selshness has been studied in several other settings, e.g. load
balancing [8, 9, 21, 29], facility location [34], and network design [3].
A simple way to model interactions between agents in a large network is with a
graphical game [19]: a graph G = (V, E) is dened whose vertices represent the
players of the game and an edge (v, w) E corresponds to the strategic interaction
between players v and w; each player v V has a nite set of strategies S
v
, which
throughout this paper will be assumed to be binary so that there are two possible
strategies for each player. A utility, or payoff, table u
v
for player v assigns a real
number u
v
(
v
,
N(v)
) to every selection of strategies by player v and the players in
vs neighborhood, i.e. the set of nodes v

such that (v, v

) E, denoted by N(v).
A pure Nash equilibrium (PNE) of the game is some state, or strategy prole, of
the game, assigning to every player v a single strategy
v
S
v
, such that no player
has a unilateral incentive to deviate. Equivalently, for every player v V ,
u
v
(
v
,
N(v)
) u
v
(

v
,
N(v)
), for every strategy

v
S
v
. (1)
When condition (1) is satised, we say that the strategy
v
is a best response to the
strategies
N(v)
.
The concept of the pure strategy Nash equilibrium is more compelling, decision
theoretically, than the concept of the mixed strategy Nash equilibriumits coun-
terpart that allows players to choose distributions over their strategy sets. This is
because it is not always meaningful in applications to assume that the players of
a game may adopt randomized strategies. Unfortunately, unlike mixed Nash equi-
libria, PNE do not always exist. It is then an important problem to study how the
existence of PNE depends on the properties of the game.
The focus of this paper is to understand how the connectivity of the under-
lying graph affects the existence of a PNE. We obtain two kinds of results. The
rst concerns the existence of a PNE in an ensemble of random graphical games
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
3
dened on a randomG(n, p)graph. We obtain a characterization of the proba-
bility that a PNE exists as a function of the density of the graph. The second set of
results concerns random graphical games on deterministic graphs. Here, we obtain
conditions on the structure of the graph under which a PNE does not exist with
high probability, suggesting also an efcient algorithm for nding witnesses of the
non-existence of a PNE. We also give complementary conditions on the structure
of the graph under which a PNE exists with constant probability. Our results are
described in detail in Section 1.3.
Comparison to Typical Constraint Satisfaction Problems. Graphical games pro-
vide a more compact way for representing large networks of interacting agents,
than normal form games, in which the game is described as if it were played on the
complete graph. Besides the compact description, one of the motivations for the in-
troduction of graphical games is their intuitive afnity to graphical statistical mod-
els; indeed, several algorithms for graphical games do have the avor of algorithms
for solving Bayes nets or constraint satisfaction problems [10, 13, 16, 22, 23].
In the other direction, the notion of a PNE provides a new genre of constraint
satisfaction problems; notably one in which, for any assignment of strategies (val-
ues) to the neighborhood of a player (variable), there is always a strategy (value)
for that player which makes the constraint (1) corresponding to that player satised
(i.e. being in best response). The reason why it might be hard to satisfy simulta-
neously the constraints corresponding to all players is the long range correlations
that may arise between players. Indeed, deciding whether a PNE exists is NP-hard
even for very sparse graphical games [16].
Viewed as a constraint satisfaction problem, the problem of the existence of
PNE poses interesting challenges. First, for natural random ensembles over payoff
tables such as the one adopted in this paper (see Denition 1.2), the expected num-
ber of PNE is 1 for any graph (this is shown for our model in the main body of
the paper; see Eq. (10)). On the contrary, for typical constraint satisfaction prob-
lems, the expected number of solutions is exponential in the size of the graph with
different exponents corresponding to different density parameters. Second, unlike
typical constraint satisfaction problems studied before, the existence of PNE is a-
priori not a monotone property of the connectivity. It is surprising that given these
novel features of the problem it is possible to obtain a result establishing a double
phase transition on the existence of PNE as described below.
1.1. Our Model. We dene the notion of a graphical game and proceed to
describe the ensemble of random graphical games studied in this paper.
DEFINITION 1.1 (Graphical Game). Given a graph G = (V, E), we dene
the neighborhood of node v V to be the set N(v) = {v

| (v, v

) E}. If S
v
is
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
4
a set associated with vertex v, for all v V , we denote by S
N(v)
:=
v

N(v)
S
v

the Cartesian product of the sets associated with the nodes in vs neighborhood.
A graphical game on G is a collection (S
v
, u
v
)
vV
, where S
v
is the strategy set
of node v and u
v
: S
v
S
N(v)
R the utility (or payoff) function (or table) of
player v. We also dene the best response function (or table) of player v to be the
function BR
v
: S
v
S
N(v)
{0, 1} such that
BR
v
(
v
,
N(v)
) = 1
v
arg
x
max{u
v
(x,
N(v)
)},
for all
v
S
v
and
N(v)
S
N(v)
.
DEFINITION 1.2 (Random Graphical Games on a Fixed Graph). Given a graph
G = (V, E) and an atomless distribution F over R, the probability distribution
D
G,F
over graphical games (S
v
, u
v
)
vV
on G is dened as follows:
S
v
= {0, 1}, for all v V ;
the payoff values {u
v
(
v
,
N(v)
)}
vV,vSv,
N(v)
S
N(v)
are mutually inde-
pendent and identically distributed according to F.
REMARK 1.3 (Invariance under Payoff Distributions). It is easy to see that the
existence of a PNE is only determined by the best response tables of the game; see
Condition (1). In particular, given that the distributions considered in this paper
are atomless, we can study PNE under D
G,F
, for any atomless F, by restricting
our attention (up to probability 0 events) to the measure D
G
over best response
tables, dened as follows
{BR
v
(0,
N(v)
)}
vV,
N(v)
S
N(v)
are mutually independent and uniform in
{0, 1};
BR
v
(1,
N(v)
) = 1 BR
v
(0,
N(v)
), for all
N(v)
S
N(v)
.
We will sometimes refer to a graphical game dened in terms of its best response
tables as an underspecied graphical game. Other times, we will overload our
terminology and just call it a graphical game. We use P
G
[.] and E
G
[.] to denote
probabilities of events and expectations respectively under the measure D
G
.
REMARK 1.4 (Invariance under Payoff Distributions II). Given our observa-
tion in Remark 1.3, it follows that it is not important to use a common distribution
F for sampling the payoffs of all the players of the game. All our results in this
work are true if different players have different distributions as long as these dis-
tributions are atomless and all payoffs values are sampled independently.
Extending the Model to Random Graphs. One of the goals of this paper is to
investigate what average degree is required in a graph for a graphical game played
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
5
on this graph to have a PNE. To study this question, it is natural to consider families
of graphs with different densities and relate the probability of PNE existence with
the density of the graph. We consider graphical games on graphs drawn from the
Erd os-R enyi, G(n, p), model, with varying values of the edge probability p. The
ensemble of graphical games we consider is formally the following.
DEFINITION 1.5. Given n N, p [0, 1] and an atomless distribution F
over R, we dene a measure D
(n,p,F)
over graphical games. A graphical game is
drawn from D
(n,p,F)
as follows:
a graph G is drawn from G(n, p);
a random graphical game is drawn from D
G,F
.
REMARK 1.6 (Invariance under Payoff Distributions III). Given our discus-
sion in Remark 1.3, it follows that in order to study PNE in the random ensemble of
Denition 1.5 it is sufcient to consider a measure that xes only the best response
tables of the players in the sampled games.
For a given n N and p [0, 1], we dene the measure D
(n,p)
over underspec-
ied graphical games. An underspecied graphical game is drawn from D
(n,p)
as
follows:
a graph G is drawn from G(n, p);
a random underspecied graphical game is drawn from D
G
.
We use P
(n,p)
[.] to denote probabilities of events under the measure D
(n,p)
and
P
G
[.] for probabilities of events measurable under G(n, p).
In the model dened in Denition 1.5 and Remark 1.6, there are two sources
of randomness: the selection of the graph, determining what players interact with
each other, and the selection of the payoff tables. Note that in the two-stage process
that samples a graphical game from our distribution, the payoff tables can only be
realized once the graph is xed. This justies the subscript G in the measure D
G
dened above.
1.2. Discussion.
Non-Monotonicity. Observe that the existence of a PNE is a non-monotone prop-
erty of p: any graphical game on the empty graph has a PNE for trivial reasons; on
the complete graph a random graphical game has a PNE with asymptotic probabil-
ity 1
1
e
(see [12, 28]); but our results indicate that, when p is in some intermediate
regime, a PNE does not exist with probability approaching 1 as n +(see The-
orem 1.10).
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
6
The non-monotonicity in the average degree of the existence of a PNE makes
the relation between PNE and connectivity non-obvious. Surprisingly, we show
(Theorem 1.9) that the convergence to a Poisson distribution of the distribution of
the number of PNE in complete graphs [26, 33] extends to much sparser graphs,
as long as the average degree is at least logarithmic in the number of players. If
the sparsity increases further, we show (Theorem 1.10) that a PNE does not exist
with high probability, while if the graph is essentially empty, PNE exist with prob-
ability 1 (Theorem 1.11). Our results establish a double phase transition consistent
with the non-monotonicity of the model.
Methodological Challenges. Our study here is an instance of studying the satis-
ability of constraint satisfaction problems (CSPs). The generic question is to inves-
tigate the effect of the structure of the constraint graph on the satisability of the
problems dened on that graph, as well as their computational complexity. In the
context of CNF formulas (corresponding to the SATisability problem) the graph
property most commonly studied in the literature is the density of the hypergraph
that contains an edge for each clause of the formula, see e.g. [14]. In other settings,
different structural properties of the constraint graph are relevant, e.g. measures of
cyclicity of the graph [6, 17]. In our case, studying the average degree reveals an
interesting, non-monotonic behavior of the model, as described above.
In a typical CSP, to show that a solution does not exist one either uses the rst
moment method to exhibit that the expected number of solutions is tiny [2], or nds
a witness of unsatisability that exists with high probability. To show that a satis-
fying assignment does exist it is quite common to use the second moment method
or its renements, which have provided some of the best bounds for satisability to
date [1]. In our model the expected number of satisfying assignments turns out to
be 1 for any graph (see Eq. (10) below). This suggests that the analysis of the prob-
lem should be harder, since in particular we cannot use the rst moment method
to establish the non-existence of a PNE. Our proof of the non-existence of PNE
(Theorems 1.10 and 1.16) uses succinct non-existence witnesses that appear with
high probability in sufciently sparse graphs. These witnesses are specic sub-
game structures that do not possess a PNE with high probability. To establish the
existence of a PNE for sufciently large densities (Theorems 1.9 and 1.13) we use
the second moment method. Further, we use Steins [4] method to establish that
the distribution of the number of PNE converges asymptotically to a Poisson(1)
distribution in this case.
1.3. Outline of Main Results. We describe rst our results for random graphs
(for the measure D
(n,p)
dened in Remark 1.6), and proceed with our results for
deterministic graphs (for the measure D
G
dened in Remark 1.3).
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
7
PNE on Random Graphs. We study how the connectivity probability p inuences
the existence of PNE for games sampled fromD
(n,p)
. The transition is described by
the following theorems applying to different levels of graph connectivity. Before
stating the theorems, we introduce some notation.
REMARK 1.7 (Order Notation). Let f(x) and g(x) be two functions dened
on some subset of the real numbers. One writes f(x) = O(g(x)) if and only if,
for sufciently large values of x, f(x) is at most a constant times g(x) in absolute
value. That is, f(x) = O(g(x)) if and only if there exists a positive real number
M and a real number x
0
such that
|f(x)| M|g(x)|, for all x > x
0
.
Similarly, we write f(x) = (g(x)) if and only if there exists a positive real num-
ber M and a real number x
0
such that
|f(x)| M|g(x)|, for all x > x
0
.
We casually use the order notation O() and () throughout the paper. When-
ever we use O(f(n)) or (f(n)) in some bound, there exists a constant c > 0
such that the bound holds true for sufciently large n if we replace the O(f(n)) or
(f(n)) in the bound by c f(n).
REMARK 1.8 (Order Notation Continued). If g(n) is a function of n N,
then we denote by (g(n)) any function f(n) such that f(n)/g(n) +, as n
+; similarly, we denote by o(g(n)) any function f(n) such that f(n)/g(n) 0,
as n +. Finally, for two functions f(n) and g(n), we write f(n) >> g(n)
whenever f(n) = (g(n)).
THEOREM 1.9 (High Connectivity). Let Z denote the number of PNE in a
graphical game sampled from D
(n,p)
, where p =
(2+) log
e
(n)
n
, = (n) > 0. For
an arbitrary constant c > 0 we assume that (n) > c and (in order for p 1)
(n)
n
log
e
(n)
2.
Under the above assumptions, for all nite n, with probability at least 1
2n
/8
over the random graph sampled from G(n, p), it holds that the total varia-
tion distance between Z and a Poisson(1) r.v. W is bounded by:
(2) ||Z W|| O(n
/8
) + exp((n)).
In other words,
(3) P
G
_
||Z W|| O(n
/8
) + exp((n))
_
1 2n
/8
.
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
8
In particular, the distribution of Z converges in total variation distance to a
Poisson(1) distribution, as n +.
(Note that the two terms on the right hand side of (2) can be of the same order
when is of the order of n/ log
e
(n).)
THEOREM 1.10 (Medium Connectivity). For all p = p(n) 1/n, if a graph-
ical game is sampled from D
(n,p)
, the probability that a PNE exists is bounded
by:
exp((n
2
p)).
For p(n) = g(n)/n, where log
e
(n)/2 > g(n) > 1, the probability that a PNE
exists is bounded by:
exp((e
log
e
(n)2g(n)
)).
In particular, the probability that a PNE exists goes to 0 as n + for all
p = p(n) satisfying
1
n
2
<< p < (0.5

(n))
log
e
(n)
n
,
where

(n) =
_
1
log
e
(n)
_
.
THEOREM 1.11 (Low Connectivity). For every constant c > 0, if a graphical
game is sampled from D
(n,p)
with p
c
n
2
, the probability that a PNE exists is at
least
_
1
c
n
2
_
n(n1)
2
e

c
2
.
Note that our upper and lower bounds for G(n, p) leave a small gap, between
p
0.5 log
e
(n)
n
and p
2 log
e
(n)
n
. The behavior of the number of PNE in this range
of p remains open. We establish the non-existence of PNE for medium connectivity
graphs via a simple structure that prevents PNE from arising, called the indifferent
matching pennies game (see Denition 1.18 below). It is natural to ask whether
our indifferent matching pennies witnesses are (similarly to isolated vertices in
connectivity) the smallest structures that prevent the existence of PNE and the last
ones to disappear.
General Graphs. We give conditions on the structure of a graph implying the
(likely) existence or non-existence of a PNE in a random game played on that
graph. The existence of a PNE is guaranteed by sufcient connectivity of the un-
derlying graph. The connectivity that we require is captured by the notion of (, )-
expansion given next.
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
9
DEFINITION 1.12 ((, )-Expansion). A graph G = (V, E) has (, )-expansion
iff every set V

such that |V

| |V | has |N(V

)| min(|V |, |V

|) neighbors.
Here we let
N(V

) = {w V : u V

with (u, w) E}.

(Note in particular that N(V

) may intersect V

).
We show the following result.
THEOREM 1.13 (Strongly Connected Graphs). Let Z denote the number of
PNE in a graphical game sampled from D
G
, where G is a graph on n vertices
that has (, )-expansion with = (1 + ) log
2
(n), =
1

and > 0. Then

the total variation distance between the distribution of Z and the distribution of a
Poisson(1) r.v. W is bounded by:
(4) ||Z W|| O(n

) +O(2
n/2
).
Next we provide a complementary condition for the non-existence of PNE. The
condition will be given in terms of the following structure.
DEFINITION 1.14 (d-Bounded Edge). An edge e = (u, v) E of a graph
G(V, E) is called d-bounded if both u and v have degrees smaller or equal to d.
We bound the probability that a PNE exists in a game sampled from D
G
as
a function of the number of d-bounded edges in G. For the stronger version of
our theorem, we also need the notion of a maximal weighted independent edge-set
dened next.
DEFINITION 1.15 (Maximal Weighted Independent Edge-Set). Given a graph
G(V, E), a subset E E of the edges is called independent if no pair of edges in
E are adjacent.
If w : E R is a function assigning weights to the edges of G, we extend w to
subsets of edges by assigning to each E E the weight w
E
=

eE
w(e). Then
we call a subset E E of edges a maximal weighted independent edge-set if E is
an independent edge-set with maximal weight among independent edge-sets.
THEOREM 1.16. A random game sampled from D
G
, where G is a graph with
at least m vertex disjoint d-bounded edges, has no PNE with probability at least
(5) 1 exp
_
m
_
1
8
_
2
2d2
_
.
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
10
In particular, if G has at least m edges that are d-bounded, then a game sampled
from D
G
has no PNE with probability at least
(6) 1 exp
_

m
2d
_
1
8
_
2
2d2
_
.
Moreover, there exists an algorithm of complexity O(n
2
+m2
d+2
) for proving that a
PNE does not exist, which has success probability given by (5) and (6) respectively.
More generally, let us assign to every edge (u, v) E the weight
w
(u,v)
:= log
e
_
1 p
(u,v)
_
,
for p
(u,v)
= 8
2
du+dv2
, where d
u
and d
v
are respectively the degrees of u and v.
Given these weights, suppose that E is a maximal weighted independent edge-set
with value w
E
. Then the probability that there exists no PNE is at least
1 exp (w
E
) .
An easy consequence of this result is that many sparse graphs, such as the line
and the grid, do not have a PNE with probability tending to 1 as the number of
players increases.
The proof of Theorem 1.16 is based on a small witness for the non-existence of
PNE, called the indifferent matching pennies game. As the name implies this game
is inspired by the simple matching pennies game. Both games are described next.
DEFINITION 1.17 (The Matching Pennies Game). We say that two players a
and b play the matching pennies game if their payoff matrices are the following, up
to permuting the players names.
Payo table of player a :
b plays 0 b plays 1
a plays 0 1 0
a plays 1 0 1
Payo table of player b :
b plays 0 b plays 1
a plays 0 0 1
a plays 1 1 0
DEFINITION 1.18 (The Indifferent Matching Pennies Game). We say that two
players a and b that are adjacent to each other in a graphical game play the in-
different matching pennies game if, for all strategy proles
N(a)N(b)\{a,b}
in the
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
11
neighborhood of a and b, the players a and b play a matching pennies game against
each other.
In other words, for all xed :=
N(a)N(b)\{a,b}
, the payoff tables of a and b
projected on
N(a)\{b}
and
N(b)\{a}
respectively are the following, up to permut-
ing the players names.
Payos to player a :
b plays 0,
other neighbors
play
N(a)\{b}
b plays 1,
other neighbors
play
N(a)\{b}
a plays 0 1 0
a plays 1 0 1
Payos to player b :
a plays 0,
other neighbors
play
N(b)\{a}
a plays 1,
other neighbors
play
N(b)\{a}
b plays 0 0 1
b plays 1 1 0
Observe that if a graphical game contains an edge (u, v) so that players u and
v play the indifferent matching pennies game then the game has no PNE. In partic-
ular, the indifferent matching pennies game provides a small witness for the non-
existence of a PNE, which is a coNP-complete problem for bounded degree graph-
ical games [16]. Our analysis implies that, with high probability over bounded
degree graphical games, there are short proofs for the non-existence of PNE which
can be found efciently. A related analysis and randomized algorithm was intro-
duced for mixed Nash equilibria in 2-player games by B ar any et al. [5].
1.4. Related Work. The number of PNE in random games with i.i.d. payoffs
has been extensively studied in the literature prior to our work: Goldberg et al. [15]
characterize the probability that a two-player random game with i.i.d. payoff tables
has a PNE, as the number of strategies tends to innity. Dresher [12] and Papavas-
silopoulos [25] generalize this result to n-player random games on the complete
graph. Powers [26] and Stanford [33] generalize the result further, showing that
the distribution of the number of PNE approaches a Poisson(1) distribution as the
number of strategies increases. Finally, Rinott et al. [28] investigate the asymptotic
distribution of PNE for a more general ensemble of random games on the com-
plete graph where there are positive or negative dependencies among the players
payoffs.
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
12
Our work generalizes the above results for i.i.d. payoffs beyond the complete
graph to random graphical games on random graphs and several families of deter-
ministic graphs. Parallel to our work, Bistra et al. [11] studied the existence of PNE
in certain families of deterministic graphs, and Hart et al. [18] obtained results for
evolutionarily stable strategies in random games. These results are related but not
directly comparable to our results.
1.5. Acknowledgement. We thank Martin Dyer for pointing out an error in a
previous formulation of Theorem 1.16. We also thank the anonymous referee for
comments that helped improve the presentation of this work.
2. Random Graphs.
2.1. High Connectivity. In this section we study the number of PNE in graph-
ical games sampled from D
(n,p)
. We show that, when the average degree is pn =
(2 + (n)) log
e
(n), where (n) > c and c > 0 is any xed constant, the distribu-
tion of the number of PNE converges to a Poisson(1) random variable, as n goes
to innity. This implies in particular that a PNE exists with probability converging
to 1
1
e
as the size of the network increases.
As in the study of the complete graph in [28], we use the following result of
Arratia et al. [4], established using Steins method. For two random variables Z, Z

supported on 0, 1, . . . we dene their total variation distance ||Z Z

|| as
||Z Z

|| :=

i=0
|Z(i) Z

(i)|.
LEMMA 2.1 ([4]). Consider arbitrary Bernoulli random variables X
i
, i =
0, . . . , N. For each i, dene some neighborhood of dependence B
i
of X
i
such that
B
i
satises that (X
j
: j B
c
i
) are independent of X
i
. Let
(7) Z =
N

i=0
X
i
, = E[Z],
and assume that > 0. Also, let
b
1
=
N

i=0

jB
i
P[X
i
= 1]P[X
j
= 1]
and b
2
=
N

i=0

jB
i
\{i}
P[X
i
= 1, X
j
= 1].
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
13
Then the total variation distance between the distribution of Z and a Poisson ran-
dom variable W

with mean is bounded by

||Z W

|| 2(b
1
+b
2
). (8)
Proof of Theorem 1.9: For ease of notation, we identify the players of the graph-
ical game with the indices 1, 2, . . . , n. We also identify pure strategy proles with
the integers in {0, . . . , 2
n
1}, mapping each integer to a strategy prole. The
mapping is dened so that, if the binary expansion of i is i(1) . . . i(n), player k
plays i(k).
Next, to each strategy prole i {0, . . . , N}, where N = 2
n
1, we assign an
indicator random variable X
i
which is 1 if the strategy prole i is a PNE. Then the
counting random variable
(9) Z =
N

i=0
X
i
corresponds to the number of PNE. Hence the existence of a PNE is equivalent to
the random variable Z being positive.
Let us condition on a realization of the graph G of the graphical game, but
not its best response tables. For a given strategy prole i, each player is in best
response with probability 1/2 over the selection of her best response table;

there-
fore E
G
[X
i
] = 2
n
, for all i, where recall that E
G
denotes expectation under the
measure D
G
. Hence, conditioning on G the expected number of PNE is
(10) E
G
[Z] = 1.
Since this holds for any realization of the graph G it follows that E[Z] = 1.
In Lemma 2.2 that follows, we characterize the neighborhood of dependence B
i
of the variable X
i
in order to be able to apply Lemma 2.1 on the collection of vari-
ables X
0
, . . . , X
N
. Note that this neighborhood depends on the graph realization,
but is independent of the realization of the payoff tables.
LEMMA 2.2. For a xed graph G, we can choose the neighborhoods of de-
pendence for the random variables X
0
, . . . , X
N
as follows:
B
0
= {j : k such that k

with (k, k

) E(G) it holds that j(k

) = 0}
and
B
i
= i B
0
= {i j : j B
0
},
where i j = (i(1) j(1), . . . , i(n) j(n)) and is the exclusive or operation.

This follows directly fromour model (Remark 1.6), following our assumption of atomless payoff
distributions (Denition 1.5).
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
14
REMARK 2.3. Intuitively, when the graph G is realized, the neighborhood of
dependence of the strategy prole 0 (variable X
0
) contains all strategy proles j
(variables X
j
) assigning 0 to all the neighbors of at least one player k. If such a
player k exists, then whether 0 or j(k) is a best response to the all-0 neighborhood
are dependent random variables (over the selection of the best response table of
player k). The denition of B
i
in terms of B
0
is justied by the symmetry of our
model.
PROOF OF LEMMA 2.2. By symmetry, it is enough to show that X
0
is inde-
pendent of {X
i
}
i / B
0
. Fix some i / B
0
. Observe that in i, each player k of the
game has at least one neighbor k

playing strategy 1. By the denition of measure

D
G
, it follows that whether strategy 0 is a best response for player k in strategy
prole 0 is independent of whether strategy i(k) is a best response for player k
in strategy prole i, since these events depend on different strategy proles of the
neighbors of k.
Now, for a xed graph G, the functions b
1
(G) and b
2
(G) (corresponding to b
1
and b
2
in Lemma 2.1) are well-dened. We proceed to bound the expectation of
these functions over the sampling of the graph G.
E
G
[b
1
(G)] = E
G
_
_
N

i=0

jB
i
P
G
[X
i
= 1]P
G
[X
j
= 1]
_
_
= E
G
_
1
(N + 1)
2
N

i=0
|B
i
|
_
=
E
G
[|B
0
|]
N + 1
; (11)
E
G
[b
2
(G)] = E
G
_
_
N

i=0

jB
i
\{i}
P
G
[X
i
= 1, X
j
= 1]
_
_
= (N + 1)

j=0
E
G
[P
G
[X
0
= 1, X
j
= 1]I[j B
0
]] . (12)
In the last line of both derivations we made use of the symmetry of the model.
Invoking symmetry again, we observe that the expectation
E
G
[P
G
[X
0
= 1, X
j
= 1]I[j B
0
]]
in (12) depends only on the number of 1s in the strategy prole j, denoted s below.
Let us write Y
s
for the indicator that the strategy prole j
s
, where the rst s players
play 1 and all the other players play 0, is a PNE. Also, write I
s
for the indicator
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
15
that this strategy is in B
0
(note that I
s
is a function of the graph only). Using this
notation, we obtain:
E
G
[b
2
(G)] = 2
n
n

s=1
_
n
s
_
E
G
[I
s
P
G
[Y
0
= 1, Y
s
= 1]]; (13)
and E
G
[b
1
(G)] = 2
n
n

s=0
_
n
s
_
E
G
[I
s
]. (14)
LEMMA 2.4. E
G
[b
1
(G)] and E
G
[b
2
(G)] are bounded as follows.
E
G
[b
1
(G)] R(n, p) :=
n

s=0
_
n
s
_
2
n
min(1, n(1 p)
s1
);
E
G
[b
2
(G)] S(n, p) :=
n

s=1
_
n
s
_
2
n
_
(1 + (1 p)
s
)
ns
(1 (1 p)
s
)
ns

.
PROOF. We begin with the study of E
G
[b
1
(G)]. Clearly, it sufces to bound
E[I
s
] by n(1 p)
s1
, for s > 0. For the strategy prole j
s
to belong in B
0
it must
be that there is at least one player who is not connected to any player in the set
S := {1, 2, . . . , s}. The probability that a specic player k is not connected to any
player in S is either (1 p)
s
or (1 p)
s1
, depending on whether k S; so it is
always at most (1 p)
s1
. By a union bound it follows that the probability there
is at least one player not connected to S is at most n(1 p)
s1
.
We now analyze E
G
[I
s
P
G
[Y
0
= 1, Y
s
= 1]]. Recall from the previous para-
graph that I
s
= 1 only when there exists a player k who is not connected to any
player in S. If there exists such a player k with the extra property that k S, then
P
G
[Y
0
= 1, Y
s
= 1] = 0, since it cannot be that both 0 and 1 are best responses for
player k when all her neighbors play 0.
Therefore the only contribution to E
G
[I
s
P
G
[Y
0
= 1, Y
s
= 1]] is from the event
every player in S is connected to at least one other player in S. Conditioning on this
event, in order for I
s
= 1 it must be that at least one of the players in S
c
:= V \ S
is not adjacent to any player in S.
Let us dene p
s
:= P
G
[ isolated node in the subgraph induced by S] and let t
denote the number of players in S
c
, which are not connected to any player in S.
Since every player outside S is non-adjacent to any player in S with probability
(1 p)
s
, the probability that exactly t players are not adjacent to S is
_
n s
t
_
[(1 p)
s
]
t
(1 (1 p)
s
)
nst
.
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
16
Moreover, conditioning on the event that exactly t players in S
c
are not adjacent to
any player in S, we have that the probability that Y
0
= 1 and Y
s
= 1 is:
1
2
t
1
2
nt
1
2
nt
.
Putting these together we obtain:
E
G
[I
s
P
G
[Y
0
= 1, Y
s
= 1]]
= p
s
ns

t=1
_
n s
t
_
[(1 p)
s
]
t
(1 (1 p)
s
)
nst
1
2
t
1
4
nt
,
=
p
s
4
n
_
(2(1 p)
s
+ (1 (1 p)
s
))
ns
(1 (1 p)
s
)
ns
_
=
p
s
4
n
_
(1 + (1 p)
s
)
ns
(1 (1 p)
s
)
ns
_
;
therefore
E
G
[b
2
(G)] =
n

s=1
2
n
_
n
s
_
p
s
_
(1 + (1 p)
s
)
ns
(1 (1 p)
s
)
ns

S(n, p).
In the appendix we show that
LEMMA 2.5.
S(n, p) O(n
/4
) + exp((n)),
and
R(n, p) O(n
/4
) + exp((n)).
Given the above bounds on E
G
[b
1
(G)] and E
G
[b
2
(G)], Markovs inequality im-
plies that with probability at least 1 n
/8
2
n
over the selection of the graph
G from G(n, p) we have
max(b
1
(G), b
2
(G)) O(n
/8
) + exp((n)). (15)
Let us condition on the event that Condition (15) holds. Under this event, Lemma 2.1
implies that:
||Z W|| 2(b
1
(G) +b
2
(G)) O(n
/8
) + exp((n))
as needed. Noting that 1 n
/8
2
n
1 2n
/8
, we obtain
(16) P
G
_
||Z W|| O(n
/8
) + exp((n))
_
1 2n
/8
.
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
17
Using the pessimistic upper bound of 2 on the total variation distance when
Condition (15) fails, we obtain
||Z W|| O(n
/8
) + exp((n)).
Taking the limit of the above bound as n +we obtain our asymptotic result.
This concludes the proof of Theorem 1.9.
2.2. Medium Connectivity.
PROOF OF THEOREM 1.10. Recall the matching pennies game from Deni-
tion 1.17. It is not hard to see that this game does not have a PNE. Hence, if a
graphical game contains two players who are connected to each other, are isolated
from all the other players, and play matching pennies against each other, then the
graphical game will have no PNE. The existence of such a witness for the non-
existence of PNE is precisely what we use to establish our result. In particular, we
show that with high probability a random game sampled from D
(n,p)
will contain
an isolated edge between two players playing a matching pennies game.
We use the following exposure argument. Label the vertices of the graph with
the integers in [n] := {1, . . . , n}. Set
1
= [n] and perform the following opera-
tions, which iteratively dene the sets of vertices
i
, i 2. If |
i
| n/2, for some
i 2, stop the process and do not proceed to iteration i:

Let j be the minimal value such that j
i
.
If j is adjacent to more than one vertex or to none, let
i+1
=
i
\ ({j}
N(j)). Go to the next iteration.
Otherwise, let j

be the unique neighbor of j. If j

has a neighbor = j, let

i+1
=
i
\ ({j, j

} N(j

)). Go to the next iteration.

Otherwise check if the players j and j

play a matching pennies game.

If
this is the case, declare NO NASH. Let
i+1
=
i
\ {j, j

}. Go to the next
iteration.
Observe that the number of vertices removed at some iteration of the process
can be upper bounded (formally, it is stochastically dominated) by
2 + Bin(n, p),

Throughout the process i represents the set of vertices that could be adjacent to an isolated
edge, given the information available to the process at the beginning of iteration i.

More precisely, check if the best response tables of the players j and j

are the same with the

best response tables of the players a and b of the matching pennies game from Denition 1.17 (up to
permutations of the players names).
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
18
where Bin(n, p) is a random variable distributed according to the Binomial distri-
bution with n trials and success probability p. This follows from the fact that the
vertices removed at some iteration of the process are either the examined vertex j
and js neighbors (the number of those is stochastically dominated by a Bin(n, p)
random variable), orif j has a single neighbor j

the removed vertices are j, j

and the neighbors of j

(the number of those is also stochastically dominated by a

Bin(n, p) random variable). Letting m := 0.02n/(np + 1), the probability that
the process runs for at most m iterations is bounded by
Pr [2m+ Bin(mn, p) n/2] exp((n)).
Condition on the information known to the exposure process up until the begin-
ning of iteration i, and assume that |
i
| > n/2. Let j be the vertex with the smallest
value in
i
. Now reveal all the neighbors of j, and if j has only one neighbor j

reveal also the neighbors of j

. The probability that j is adjacent to a node j

who
has no other neighbors is at least
n
2
p(1p)
2n
=: p
iso
; note that we made use of the
condition |
i
| > n/2 in this calculation. Conditioning on this event, the probability
(over the selection of the payoff tables) that j and j

play a matching pennies game

is
1
8
=: p
mp
. Hence, the probability of outputting NO NASH in iteration i is at least
1
8
1
2
np(1 p)
2n
=: p
imp
.
The probability that the game has a PNE is upper bounded by the probability
that the process described above does not return NO NASH, at any point through
its completion. To upper bound the latter probability, let us imagine the following
alternative process:
1. Stage 1: Toss n coins independently at random with head probability p
iso
.
Let I
1
, I
2
, . . . , I
n
{0, 1}, where 1 represents heads and 0 represents
tails, be the outcomes of these coin tosses.
2. Stage 2: Toss n coins independently at random with head probability p
mp
.
Let M
1
, M
2
, . . . , M
n
{0, 1}, be the outcomes of these coin tosses.
3. Stage 3: Run through the exposure process in the following way. At each
iteration i:
conditioning on the information available to the exposure process at the
beginning of the iteration, compute the probability p
j
that the vertex j
corresponding to the smallest number in
i
is adjacent to an isolated
edge; given the discussion above it must be that p
j
p
iso
;
if I
i
= 1, then create an isolated edge connecting the player j to a
random vertex j

i
\ {j}, forbidding all other edges from j or j

to
any other player, and make the players j and j

play a matching pennies

game if M
i
= 1; if they do output NO NASH;
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
19
if I
i
= 0, then sample the neighborhood of j from the following mod-
ied model:
with probability
p
j
p
iso
1p
iso
, create an isolated edge connecting the
player j to a random vertex j

i
\ {j}, forbidding all other
edges from j or j

to any other player, and make the players j and

j

play a matching pennies game with probability p

mp
; if both of
these happen, output NO NASH;
with the remaining probability, sample the neighborhood of j and
the neighborhood of the potential unique neighbor j

fromG(n, p),
conditioning on j not being adjacent to an isolated edge.
Dene
i+1
from
i
appropriately and exit the process if |
i+1
| n/2.
It is clear that the process given above can be coupled with the process dened
earlier to exhibit the same behavior. But it is easier to analyze. In particular, letting
S :=

m
i=1
I
i
M
i
, the probability that a Nash equilibrium does not exists can be
lower bounded as follows:
P
G
[ a PNE] Pr
_
S 1
process runs for
at least m steps
_
Pr [S 1] Pr
_
process runs for
less than m steps
_
1 (1 p
imp
)
m
exp((n)).
Hence, the probability that a PNE exists can be upper bounded by
exp((n)) +
_
1
1
16
np(1 p)
2n
_
m
exp((n)) + exp((mnp(1 p)
2n
))
exp((mnp(1 p)
2n
)).
For p 1/n the last expression is
exp((n
2
p)),
while for p = g(n)/n where g(n) 1 the expression is
exp((n(1 p)
2n
)) = exp((ne
2g(n)
)) = exp((e
log
e
(n)2g(n)
)).
This concludes the proof of Theorem 1.10.
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
20
2.3. Low Connectivity.
PROOF OF THEOREM 1.11. Note that if the graphical game is comprised of
isolated edges that are not matching pennies games then a PNE exists. (This can be
checked easily by enumerating all best response tables for a 2 2 game.) We wish
to lower bound the probability of this event. To do this, it is convenient to sample
the graphical game in two stages as follows: At the rst stage we decide for each of
the possible
_
n
2
_
edges whether the edge is present (with probability p) and whether
it is predisposed to be a matching pennies game (independently with probability
1/8); by predisposed we mean that the edge will be set to be a matching pennies
game if the edge turns out to be isolated. At the second stage, we do the following:
for an edge that is both isolated and predisposed, we assign random payoff tables to
its endpoints conditioning on the resulting game being a matching pennies game;
for an isolated edge that is not predisposed, we assign random payoff tables to its
endpoints conditioning on the resulting game not being a matching pennies game;
nally, for any node that is part of a connected component with 0 or at least 2 edges
we assign random payoff tables to the node. The probability that there is no edge
in the rst stage that is both present and predisposed is
(1 p/8)
(
n
2
)
.
Conditioning on this event, all present edges are not predisposed. Note also that,
when c is xed, the probability that there exists a pair of adjacent edges is o(1). It
follows that the probability that all present edges are not predisposed and no pair
of edges intersect can be lower bounded as
(1 p/8)
(
n
2
)
o(1) =
_
1
c
8n
2
_
n(n1)
2
o(1).
But, as explained above if all edges are isolated and none of them is a matching
pennies game a PNE exists. Hence, the probability that a PNE exists is at least
_
1
c
8n
2
_
n(n1)
2
o(1) e

c
16
.
3. Deterministic Graphs.
3.1. A Sufcient Condition for Existence of Equilibria: Strong Connectivity.
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
21
PROOF OF THEOREM 1.13. We use the same notation as in the proof of Theo-
rem 1.9, except that we make the slight modication of setting N := 2
n
1. Recall
that X
i
, i = 0, 1, . . . , N 1, is the indicator random variable of the event that the
strategy prole encoded by the number i is a PNE. It is rather straightforward (see
the proof of Theorem 1.9) to show that
E[Z] = E
_
N1

i=0
X
i
_
= 1.
As in the proof of Theorem 1.9, to establish our result, it sufces to bound the
following quantities.
b
1
(G) =
N1

i=0

jB
i
P[X
i
= 1]P[X
j
= 1],
b
2
(G) =
N1

i=0

jB
i
\{i}
P[X
i
= 1, X
j
= 1],
where the neighborhoods of dependence B
i
are dened as in Lemma 2.2. For S
{1, . . . , n}, denote by i(S) the strategy prole in which the players of the set S
play 1 and the players not in S play 0. Then writing 1(j B) for the indicator of
the event that j B we have:
b
2
(G) =
N1

i=0

jB
i
\{i}
P[X
i
= 1, X
j
= 1]
=
N1

i=0

j=i
P[X
i
= 1, X
j
= 1]1(j B
i
)
= N

j=0
P[X
0
= 1, X
j
= 1]1(j B
0
) (by symmetry)
= N
n

k=1

S,|S|=k
P[X
0
= 1, X
i(S)
= 1]1(i(S) B
0
).
We will bound the sum above by bounding
(17) N
n

k=1

S,|S|=k
P[X
0
= 1, X
i(S)
= 1]1(i(S) B
0
),
and
(18) N
n

k=n+1

S,|S|=k
P[X
0
= 1, X
i(S)
= 1]1(i(S) B
0
)
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
22
separately.
Note that if some set S satises |S| n then |N(S)| |S| since the graph
has (, )-expansion. Moreover, each vertex (player) of the set N(S) is playing its
best response to the strategies of its neighbors in both proles 0 and i(S) with
probability
1
4
, since its environment is different in the two proles. On the other
hand, each player not in that set is in best response in both proles 0 and i(S) with
probability at most
1
2
. Hence, we can bound (17) by
N
n

k=1

S,|S|=k
P[X
0
= 1, X
i(S)
= 1]
N
n

k=1

S,|S|=k
_
1
2
_
nk
_
1
4
_
k
=
n

k=1
_
n
k
_
_
1
2
_
k
<
_
1 +
_
1
2
_

_
n
1 en

.
To bound the second term, notice that, if some set S satises |S| n + 1, then
since the graph has (, )-expansion N(S) V and, therefore, the environment of
every player is different in the two proles 0 and i(S). Hence, 1(i(S) B
0
) = 0.
By combining the above we get that
b
2
(G) en

.
It remains to bound the expression b
1
(G). We have
b
1
(G) 2
n
=
N1

i=0

jB
i
\{i}
P[X
i
= 1]P[X
j
= 1]
=
N1

i=0

j=i
P[X
i
= 1]P[X
j
= 1]1(j B
i
)
= 2
n

j=0
1(j B
0
)
= 2
n
n

k=1

S,|S|=k
1(i(s) B
0
) + 2
n
n

k=n+1

S,|S|=k
1(i(s) B
0
).
The second term is zero as before. For all large enough n the rst summation
contains at most 2
n/2
terms and is therefore bounded by 2
n/2
. It follows that
b
1
(G) +b
2
(G) en

+ 2
n/2
.
An application of the result by Arratia et al. [4] concludes the proof of Theo-
rem 1.13.
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
23
3.2. ASufcient Condition for the Non-Existence of Equilibria: Indifferent Match-
ing Pennies. In this section we provide a proof of Theorem 1.16. Recall that an
edge of a graph is called d-bounded if both adjacent vertices have degrees smaller
or equal to d. Theorem 1.16 species that any graph with many such edges is un-
likely to have PNE. We proceed to the proof of the claim.
PROOF OF THEOREM 1.16. Consider a d-bounded edge in a game connect-
ing two players a and b; suppose that each of these players interacts with d 1
(or fewer) other players denoted by a
1
, a
2
. . . a
d1
and b
1
, b
2
. . . b
d1
.

Recall that
if a and b play an indifferent matching pennies game against each other then the
game has no PNE. The key observation is that a d-bounded edge is an indifferent
matching pennies game with probability at least (
1
8
)
2
2d2
=: p
imp
since a ran-
dom two-player game is a matching pennies game with probability
1
8
and there
are at most 2
2d2
possible pure strategy proles for the players a
1
, a
2
. . . a
d1
,
b
1
, b
2
. . . b
d1
; for each of these pure strategy proles the game between a and b
must be a matching pennies game.
For a collection of m vertex disjoint edges, observe that the events that each of
them is an indifferent matching pennies game are independent. Hence, the prob-
ability that the game has a PNE is upper bounded by the probability that none of
these edges is an indifferent matching pennies game, which is upper bounded by
(1 p
imp
)
m
exp(mp
imp
) = exp
_
m
_
1
8
_
2
2d2
_
.
For the second claim of the theorem note that, if there are m d-bounded edges,
then there must be at least m/(2d) vertex disjoint d-bounded edges.
The algorithmic statement follows from the fact that we may nd all nodes
with degree d in time O(n
2
), and then nd all edges joining two such nodes
in another O(n
2
) time, with the use of the appropriate data structures; these edges
are the d-bounded edges of the graph. Then in time O(m2
d+2
) we can check if the
endpoints of any such edge play an indifferent matching pennies game.
The nal claim of the theorem has a similar proof where now the potential
witnesses for the non-existence of a PNE are the edges in E.
Many random graphical games on deterministic graphs such as players arranged
on a line, grid, or any other bounded degree graph (with (1) edges) are special
cases of the above theorem and hence are unlikely to have PNE asymptotically.

We allow these lists to share players.

imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
24
References.
[1] D. Achlioptas and Y. Peres. The threshold for random k-sat is 2k (ln 2 - o(k)). In STOC
03: Proceedings of the thirty-fth annual ACM symposium on Theory of computing, pages
223231, New York, NY, USA, 2003. ACM Press.
[2] N. Alon and J. H. Spencer. The probabilistic method. Wiley-Interscience Series in Discrete
Mathematics and Optimization. Wiley-Interscience [John Wiley & Sons], New York, second
edition, 2000. With an appendix on the life and work of Paul Erd os.
[3] E. Anshelevich, A. Dasgupta, J. Kleinberg, E. Tardos, T. Wexler, and T. Roughgarden. The
price of stability for network design with fair cost allocation. In FOCS 04: Proceedings of the
45th Symposium on Foundations of Computer Science, 2004.
[4] R. Arratia, L. Goldstein, and L. Gordon. Two moments sufce for poisson approximations:
The chen-stein method. The Annals of Probability, 17:925, 1989.
[5] I. B ar any, S. Vempala, and A. Vetta. Nash equilibria in random games. In FOCS 05: Pro-
ceedings of the of the 46th Symposium on Foundations of Computer Science, pages 134145,
2005.
[6] C. Beeri, R. Fagin, D. Maier, and M. Yannakakis. On the desirability of acyclic database
schemes. J. ACM, 30(3):479513, 1983.
[7] J. R. Correa, A. S. Schulz, and N. E. Stier-Moses. Selsh routing in capacitated networks.
Mathematics of Operations Research, 29(4):961976, Nov. 2004.
[8] A. Czumaj, P. Krysta, and B. V ocking. Selsh trafc allocation for server farms. In STOC 02:
Proceedings of the 34th ACM Symposium on the Theory of Computing, pages 287296, 2002.
[9] A. Czumaj and B. V ocking. Tight bounds for worst-case equilibria. In SODA 02: Proceedings
of the 13th ACM-SIAM Symposium on Discrete algorithms, pages 413420, 2002.
[10] C. Daskalakis and C. H. Papadimitriou. Computing pure nash equilibria in graphical games
via markov random elds. In EC 06: Proceedings of the 7th ACM conference on Electronic
commerce, pages 9199, New York, NY, USA, 2006. ACM Press.
[11] B. Dilkina, C. P. Gomes, and A. Sabharwal. The impact of network topology on pure nash
equilibria in graphical games. In AAAI-07: Proceedings of the 22nd Conference on Articial
Intelligence, pages 4249, Vancouver, BC, 2007.
[12] M. Dresher. Probability of a pure equilibrium point in n-person games. Journal of Combinato-
rial Theory, 8:134145, 1970.
[13] E. Elkind, L. A. Goldberg, and P. Goldberg. Nash equilibria in graphical games on trees re-
visited. In EC 06: Proceedings of the 7th ACM conference on Electronic commerce, pages
100109, New York, NY, USA, 2006. ACM Press.
[14] E. Friedgut. Sharp thresholds of graph properties, and the k-sat problem. J. Amer. Math. Soc.,
12(4):10171054, 1999. With an appendix by Jean Bourgain.
[15] K. Goldberg, A. Goldman, and M. Newman. The probability of an equilibrium point. Journal
of Research of the National Bureau of Standards, 72B:93101, 1968.
[16] G. Gottlob, G. Greco, and F. Scarcello. Pure nash equilibria: hard and easy games. In TARK
03: Proceedings of the 9th conference on Theoretical aspects of rationality and knowledge,
pages 215230, New York, NY, USA, 2003. ACM Press.
[17] G. Gottlob, N. Leone, and F. Scarcello. Hypertree decompositions and tractable queries. J.
Comput. Syst. Sci., 64(3):579627, 2002.
[18] S. Hart, Y. Rinott, and B. Weiss. Evolutionarily stable strategies of random games, and the
vertices of random polygons. Annals of Applied Probability, 18:259287, 2008.
[19] M. J. Kearns, M. L. Littman, and S. P. Singh. Graphical models for game theory. In UAI 01:
Proceedings of the 17th Conference in Uncertainty in Articial Intelligence, pages 253260,
San Francisco, CA, USA, 2001. Morgan Kaufmann Publishers Inc.
[20] J. Kleinberg. The emerging intersection of social and technological networks: Open questions
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
25
and algorithmic challenges. In FOCS 06: Proceedings of the of the 47th Symposium on Foun-
dations of Computer Science, 2006.
[21] E. Koutsoupias and C. Papadimitriou. Worst-case equilibria. In STACS 99: Proceedings of
the 16th Annual Symposium on Theoretical Aspects of Computer Science, volume 1563, pages
404413, 1999.
[22] M. L. Littman, M. J. Kearns, and S. P. Singh. An efcient, exact algorithm for solving tree-
structured graphical games. In NIPS, pages 817823, 2001.
[23] L. E. Ortiz and M. J. Kearns. Nash propagation for loopy graphical games. In NIPS, pages
793800, 2002.
[24] C. H. Papadimitriou. Algorithms, games, and the internet. In STOC 01: Proceedings of the
33rd Annual ACM Symposium on Theory of Computing, pages 749753, 2001.
[25] G. P. Papavassilopoulos. On the probability of existence of pure equilibria in matrix games.
Journal of Optimization Theory and Applications, 87:419439, 1995.
[26] I. Y. Powers. Limiting distributions of the number of pure strategy nash equilibria in n-person
games. International Journal of Game Theory, 19:277286, 1990.
[27] P. Raghavan. The changing face of web search: algorithms, auctions and advertising. In STOC
06: Proceedings of the 38th Annual ACM Symposium on Theory of Computing, page 129,
2006.
[28] Y. Rinott and M. Scarsini. On the number of pure strategy nash equilibria in random games.
Games and Economic Behavior, 33:274293, 2000.
[29] T. Roughgarden. Stackelberg scheduling strategies. In STOC 01: Proceedings of the 33th
ACM Symposium on Theory of Computing, pages 104113, 2001.
[30] T. Roughgarden. The price of anarchy is independent of the network topology. In STOC 02:
Proceedings of the 34th ACM Symposium on the Theory of Computing, pages 428437, 2002.
[31] T. Roughgarden and E. Tardos. How bad is selsh routing? J. ACM, 49(2):236259, 2002.
[32] T. Roughgarden and E. Tardos. Bounding the inefciency of equilibria in nonatomic congestion
games. Games and Economic Behavior, 47(2):389403, 2004.
[33] W. Stanford. A note on the probability of k pure nash equilibria in matrix games. Games and
Economic Behavior, 9:238246, 1995.
[34] A. Vetta. Nash equilibria in competitive societies with applications to facility location. In
FOCS 02: Proceedings of the 43rd Symposium on Foundations of Computer Science, pages
416425, 2002.
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
26
APPENDIX A: OMITTED PROOFS
PROOF OF LEMMA 2.5. We need to bound the functions S(n, p) and R(n, p).
We begin with S.
Bounding S. Recall that
S(n, p) :=
n

s=1
_
n
s
_
2
n
_
(1 + (1 p)
s
)
ns
(1 (1 p)
s
)
ns

.
We split the range of the summation into four regions and bound the sum over each
region separately. We begin by choosing = () as follows
(i) if
1790
105
, we choose =
_

2+
_
20
;
(ii) if >
1790
105
, we choose =

2+
.
Given our choice of = () we dene the following regions in the range of s
(wheredepending on Regions I and/or III may be empty and Region IV may
have overlap with Region II):
I. {s N | 1 s <

(2+)p
};
II. {s N |

(2+)p
s < n};
III. {s N | n s <
1
2+
n};
IV. {s N |
1
2+
n s < n}.
We then write
S(n, p) S
I
(n, p) +S
II
(n, p) +S
III
(n, p) +S
IV
(n, p),
where S
I
(n, p) denotes the sum over region I etc., and bound each term separately.
Region I. The following lemma will be useful.
LEMMA A.1. For all > 0, p (0, 1) and s such that 1 s <

(2+)p
,
(1 p)
s
1
(2 + 0.5)sp
2 +
.
PROOF. First note that, for all k 1,
_
s
2k + 2
_
p
2k+2

_
s
2k + 1
_
p
2k+1
. (19)
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
27
To verify the latter note that it is equivalent to
s 2k + 1 +
2k + 2
p
,
which is true since s

(2+)p
=
1
(
2

+1)p

1
p
.
Using (19), it follows that
(1 p)
s
1
_
s
1
_
p +
_
s
2
_
p
2
. (20)
Note nally that
0.5
2 +
sp >
s(s 1)
2
p
2
,
which applied to (20) gives
(1 p)
s
1
(2 + 0.5)sp
2 +
.
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
28
Assuming that Region I is non-empty and applying Lemma A.1 we get:
S
I
(n, p)

s<

(2+)p
_
n
s
_
2
n
(1 + (1 p)
s
)
ns

s<

(2+)p
_
n
s
_
2
n
_
1 + 1
(2 + 0.5)sp
2 +
_
ns

s<

(2+)p
_
n
s
_
2
s
_
1
(1 + 0.25)sp
2 +
_
ns

s<

(2+)p
_
n
s
_
2
s
exp
_

(1 + 0.25)sp
2 +
(n s)
_

s<

(2+)p
_
n
s
_
2
s
exp
_

(1 + 0.25)sp
2 +
n
_
exp
_
(1 + 0.25)sp
2 +
s
_

s<

(2+)p
_
n
s
_
2
s
exp ((1 + 0.25) log
e
(n) s) exp
_
(1 + 0.25)
(2 +)
2
s
_

s<

(2+)p
n
s
2
s
n
(1+0.25)s
exp
_
1
2
s
_

s<

(2+)p
_

e
2
_
s
n
0.25s

s<

(2+)p
_

e
2
_
s
n
0.25
n
0.25

s<
2
(2+)p
_

e
2
_
s
= O(n
0.25
)
_
since

e
2
< 1
_
.
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
29
Region II. We have
S
II
(n, p)

(2+)p
s<n
_
n
s
_
2
n
(1 + (1 p)
s
)
n

(2+)p
s<n
_
n
s
_
2
n
_
1 +e
ps
_
n

(2+)p
s<n
_
n
n
_
2
n
_
1 +e
p

(2+)p
_
n
n
_
n
n
__
1 +e

2+
2
_
n
n2
nH()
(n + 1)
_
1 +e

2+
2
_
n
n(n + 1)
_
2
H()

1 +e

2+
2
_
n
. (21)
In the above derivation H() represents the entropy function, and for the second to
last derivation we used the fact that:
_
n
k
_
(n + 1)2
nH(
k
n
)
. (22)
Our denition of the function = () guarantees that when
1790
105
:
_
2
H()

1 +e

2+
2
_
0.999,
while when >
1790
105
:
_
2
H()

1 +e

2+
2
_
0.99.
Using the above and (21) we obtain
S
II
(n, p) = exp((n)). (23)
Region III. Let us assume that the region is non-empty. We show that each
positive term in the summation S
III
(n, p) is exponentially small. Since there are
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
30
O(n) terms in the summation it follows then that S
III
(n, p) is exponentially small.
_
n
s
_
2
n
(1 + (1 p)
s
)
n

_
n
s
_
2
n
(1 +e
ps
)
n

_
n
s
_
2
n
(1 +e
pn
)
n

_
n
s
_
2
n
_
1 +e
(2+)log
e
(n)
_
n
=
_
n
s
_
2
n
_
1 +
1
n
(2+)
_
n
=
_
n
s
_
2
n
_
1 +
1
n
(2+)
_
n
(2+)
n
1(2+)

_
n
s
_
2
n
e
n
1(2+)
(n + 1)2
nH(
s
n
)
2
n
e
n
1(2+)
(n + 1)2
nH(
1
2+
)
2
n
e
n
1(2+)
= (n + 1)2
n(H(
1
2+
)1)
e
n
1(2+)
, (24)
where in the third-to-last line of the derivation we employed the bound of Equa-
tion (22). Notice that the RHS of (24), seen as a function of > 0 and > 0, is
decreasing in both. Since > c, our choice of = () implies that >
_
c
c+2
_
20
.
Hence, we can bound the RHS of (24) as follows:
(n + 1)2
n(1H(
1
2+c
))
e
n
1(2+c)
(
c
c+2
)
20
= exp((n)),
where we used the fact that c is a constant, and therefore the factor e
n
1(2+c)
(
c
c+2
)
20
is sub-exponential in n, while the factor 2
n(1H(
1
2+c
))
is exponentially small in
n.
Region IV. Note that, if xk 1, then by the mean value theorem
(1 +x)
k
(1 x)
k
2x max
11/ky1+1/k
ky
k1
= 2kx(1 + 1/k)
k1
2ekx.
We can apply this for k = n s and x = (1 p)
s
since
(n s)(1 p)
s
(n s)e
ps
(n s)e

(2+) log
e
(n)
n
n
2+

n s
n
1.
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
31
Hence, S
IV
(n, p) is bounded as follows.
S
IV
(n, p)

n
2+
sn
_
n
s
_
2
n
2e(n s)(1 p)
s
2e 2
n
n

n
2+
sn
_
n
s
_
(1 p)
s
2e 2
n
n(1 + (1 p))
n
2en
_
1
p
2
_
n
2ene

p
2
n
2ene

(2+) log
e
(n)
2n
n
2enn

2+
2
2en

2
.
Putting everything together. Combining the above we get that
S(n, p) O(n
/4
) + exp((n)).
Bounding R. Observe that
R(n, p) = 2
n
+
n

s=1
_
n
s
_
2
n
min(1, n(1 p)
s1
).
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010
32
We bound R as follows.
R(n, p) 2
n

s=1
_
n
s
_
2
n
min(1, nexp(p(s 1)))
2
n

1s
n
6+3
3+
_
n
s
_
+ 2
n

s>
n
6+3
3+
_
n
s
_
nexp(p(s 1))
2
n

1s
n
6+3
3+
(n + 1)2
nH(s/n)
+ 2
n

s>
n
6+3
3+
_
n
s
_
nexp(p(s 1))
n(n + 1)2
n
2
nH(
3+
6+3
)
+ 2
n

s>
n
6+3
3+
_
n
s
_
nexp(p(s 1))
exp((n)) + 2
n

s>
n
6+3
3+
_
n
s
_
nexp(p(s 1)),
where in the last line of the derivation we used that > c > 0 for some absolute
constant c. To bound the last sum we observe that when s >
n
6+3
3+
we have
nexp(p(s 1)) nexp
_

(2 +) log
e
(n)
n
_
n
6+3
3+
1
__
n n

2+
6+3
3+
exp
_
(2 +) log
e
(n)
n
_
n
/3
n
2/n
n
/n
= O(n
/4
).
Using this bound and the fact

n
s=0
_
n
s
_
= 2
n
concludes the proof.
ADDRESS OF THE FIRST AUTHOR
COMPUTER SCIENCE AND ARTIFICIAL INTELLIGENCE LABORATORY
DEPARTMENT OF ELECTRICAL ENGINEERING AND COMPUTER SCIENCE
MIT, CAMBRIDGE, MA 02139
E-MAIL: [email protected]
ADDRESS OF THE SECOND AUTHOR
DEPARTMENT OF ELECTRICAL ENGINEERING - SYSTEMS
UNIVERSITY OF SOUTHERN CALIFORNIA
LOS ANGELES, CA 90089
E-MAIL: [email protected]
ADDRESS OF THE THIRD AUTHOR
DEPARTMENT OF STATISTICS,
UC BERKELEY
BERKELEY, CA 94720
E-MAIL: [email protected]
imsart-aap ver. 2007/12/10 file: RandomGamesAAP_ma14.tex date: May 17, 2010