# Simple strategies for large zero-sum games with applications to complexity theory

@inproceedings{Lipton1994SimpleSF, title={Simple strategies for large zero-sum games with applications to complexity theory}, author={Richard J. Lipton and N. Young}, booktitle={STOC '94}, year={1994} }

Von Neumann's Min-Max Theorem guarantees that each player of a zero-sum matrix game has an optimal mixed strategy. This paper gives an elementary proof that each player has a near-optimal mixed strategy that chooses uniformly at random from a multiset of pure strategies of size logarithmic in the number of pure strategies available to the opponent.
For exponentially large games, for which even representing an optimal mixed strategy can require exponential space, it follows that there are near… Expand

#### Topics from this paper

#### Paper Mentions

#### 92 Citations

Randomized sampling for large zero-sum games

- Computer Science
- 49th IEEE Conference on Decision and Control (CDC)
- 2010

A procedure by which a player can compute mixed policies that, with high probability, are security policies against an adversary that is also using randomized methods to solve the game is provided. Expand

Playing large games using simple strategies

- Mathematics, Computer Science
- EC '03
- 2003

The existence of ε-Nash equilibrium strategies with support logarithmic in the number of pure strategies is proved and it is proved that if the payoff matrices of a two person game have low rank then the game has an exact Nash equilibrium with small support. Expand

On the Complexity of Succinct Zero-Sum Games

- Mathematics, Computer Science
- Computational Complexity Conference
- 2005

It is proved that approximating the value of a succinct zero-sum game to within an additive factor is complete for the class promise-S/sub 2//sup p/, the "promise" version of S/ sub 2//Sup p/, which is the first natural problem shown complete for this class. Expand

A Game-theoretic Classiication of Interactive Complexity Classes (extended Abstract)

- 1995

Game-theoretic characterizations of complexity classes have often proved useful in understanding the power and limitations of these classes. One well-known example tells us that PSPACE can be… Expand

On the Complexity of Succinct Zero-Sum Games

- Computer Science, Mathematics
- 20th Annual IEEE Conference on Computational Complexity (CCC'05)
- 2005

It is proved that approximating the value of a succinct zero-sum game to within an additive error is complete for the class promise-S p 2, the “promise” version of S^{p}_{2}, and a ZPPNP algorithm is described for constructing approximately optimal strategies, and hence for approximates the value, of a given succinctZero-Sum game. Expand

Randomized sampling for large zero-sum games

- Mathematics, Computer Science
- Autom.
- 2013

This paper formalizes a procedure, termed as the sampled security policy (SSP) algorithm, by which a player can compute policies that, with a high confidence, are security policies against an adversary using randomized methods to explore the possible outcomes of the game. Expand

Simple approximate equilibria in large games

- Computer Science, Mathematics
- EC
- 2014

It is proved that in every normal form n-player game with m actions for each player, there exists an approximate Nash equilibrium in which each player randomizes uniformly among a set of O(log m + log n) pure actions, and an inverse connection between the entropy of Nash equilibria in the game, and the time it takes to find such an approximation using the random sampling method is established. Expand

Approximating the Minmax Value of Three-Player Games within a Constant is as Hard as Detecting Planted Cliques

- Computer Science, Mathematics
- SAGT
- 2012

Two reductions are presented: a randomised many-one reduction and a deterministic Turing reduction, which may be seen as a derandomisation of the former and may be used to argue for hardness of approximating the minmax value based on a hardness assumption about deterministic algorithms. Expand

A game-theoretic classification of interactive complexity classes

- Computer Science
- Proceedings of Structure in Complexity Theory. Tenth Annual IEEE Conference
- 1995

This paper uses polynomially definable game systems to derive new characterizations of the complexity classes NEXP and coNEXP, and investigates the connection between game theory and interactive computation. Expand

2 5 D ec 2 01 7 The Query Complexity of Correlated Equilibria ∗ †

- 2018

We consider the complexity of finding a correlated equilibrium of an n-player game in a model that allows the algorithm to make queries on players’ payoffs at pure strategy profiles. Randomized… Expand

#### References

SHOWING 1-10 OF 36 REFERENCES

Lower bounds by probabilistic arguments

- Mathematics, Computer Science
- 24th Annual Symposium on Foundations of Computer Science (sfcs 1983)
- 1983

It is proved that, to compute the majority function of n Boolean variables, the size of any depth-3 monotone circuit must be greater than 2nε, and thesize of any width-2 branching program must have super-polynomial growth. Expand

A theory of learning simple concepts under simple distributions and average case complexity for the universal distribution

- Computer Science
- 30th Annual Symposium on Foundations of Computer Science
- 1989

It is shown that for essentially all algorithms, if the inputs are distributed according to the universal distribution, then the average- case complexity is of the same order of magnitude as the worst-case complexity. Expand

Complete and incomplete randomized NP problems

- Computer Science
- 28th Annual Symposium on Foundations of Computer Science (sfcs 1987)
- 1987

Leonid Levin generalized the NP completeness theory to the case of properly defined randomized NP problems and proved the completeness of a randomized version of the bounded tiling problem with respect to (appropriately generalized) Ptime reductions. Expand

On the theory of average case complexity

- Mathematics, Computer Science
- STOC '89
- 1989

The equivalence of search and decision problems in the context of average case complexity; an initial analysis of the structure of distributional-NP under reductions which preserve average polynomial-time; definitions and basic theorems regarding other complexity classes such as average log-space. Expand

On sparse approximations to randomized strategies and convex combinations

- Mathematics
- 1994

Abstract A randomized strategy or a convex combination may be represented by a probability vector p = (p1, …, pm). p is called sparse if it has few positive entries. This paper presents an… Expand

Greedy Algorithms by Derandomizing Unknown Distributions

- Computer Science
- 1994

Simple strategies for large zero-sum games with applications to complexity theory and Primal-dual RNC approximation algorithms for (multi)-set (multi-cover and covering integer programs. Expand

A Graph-Theoretic Game and Its Application to the k-Server Problem

- Computer Science, Mathematics
- SIAM J. Comput.
- 1995

The main result is that the value of the game on any $n$-vertex graph is bounded above by $\exp(O(\sqrt{\log n \log\log n}))", which has potential application to the design of communication networks. Expand

The strength of weak learnability

- Computer Science
- Mach. Learn.
- 1990

In this paper, it is shown that the two notions of learnability are equivalent, and a method is described for converting a weak learning algorithm into one that achieves arbitrarily high accuracy. Expand

Probability Inequalities for sums of Bounded Random Variables

- Mathematics
- 1963

Abstract Upper bounds are derived for the probability that the sum S of n independent random variables exceeds its mean ES by a positive number nt. It is assumed that the range of each summand of S… Expand

No better ways to generate hard NP instances than picking uniformly at random

- Computer Science, Mathematics
- Proceedings [1990] 31st Annual Symposium on Foundations of Computer Science
- 1990

It is shown that every DNP problem complete for P-time computable distributions is also complete for all distributions that can be sampled, making the concept of average-case NP completeness robust and the question of the average- case complexity of complete DNP problems a natural alternative to P=?NP. Expand