Index and Robustness of Mixed Equilibria: An Algebraic Approach
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Economics. The day in economics: theory, mechanisms and the evidence beneath them.
Ruth: I'm Ruth.
Warren: And I'm Warren, here with Iris.
Ruth: We're looking at the paper: "Index and Robustness of Mixed Equilibria".
Warren: Here's the gist of the paper. Index and Robustness of Mixed Equilibria: An Algebraic Approach presents a new method based on algebraic properties of polynomial systems to compute the index of equilibria in finite games, revealing that any integer can be the index of an isolated completely mixed equilibrium, and establishing a sharp tool for verifying payoff-robustness.
Ruth: The Core Results Any integer can be the index of an isolated completely mixed equilibrium of a finite game.
Warren: In the class called monogenic, the index can be 0, +1 or −1 only.
Ruth: Non-zero index is equivalent to payoff-robustness in this class.
Title and authors: Ruth: So we're looking at this paper called "Index and Robustness of Mixed Equilibria: An Algebraic Approach," and what it's doing is focusing on the index of these mixed equilibria in finite games. Basically, it’s about giving us a new algebraic way to figure out how stable these equilibria are.
Warren: It sounds like they're trying to move beyond just seeing if an equilibrium exists and instead quantifying its robustness using this index number.
Iris: So, the core idea here seems to be linking the index of an equilibrium point directly to whether it's payoff-robust or not, which is a big question in game theory.
Ruth: Exactly. They are showing that any integer can actually be the index for an isolated completely mixed equilibrium in a finite game, which is pretty broad coverage.
Warren: That’s interesting because usually you only see simple things like one or zero when you look at these fixed-point problems and their behavior around the equilibrium point.
Iris: The paper seems to be tackling that general case by developing this new method based on algebraic properties of the polynomial system that describes those completely mixed equilibria.
The paper's summary: Ruth: They summarize this approach by saying they developed a new method for index computation based on algebraic invariants of the polynomial system describing completely mixed equilibria, and they stress that this method is perturbation-free, which is huge because standard formulas often break down in singular cases.
Warren: So, instead of relying on those tricky standard formulas that fail when things get weird, they've built a computational tool that works even in those difficult spots.
Iris: And the main result they highlight is a converse to an existing result: if you have an isolated and completely mixed equilibrium where the Jacobian has corank one at that point, then the equilibrium has index zero if and only if it isn't payoff-robust.
Ruth: That’s a sharp tool because it lets us check for payoff-robustness by just checking the index, which is a significant simplification for verification work.
Warren: It simplifies things because you don't need to run complex perturbation tests on payoffs; you can use this algebraic index computation instead.
Iris: They also extend this idea to extensive form games, proposing a way to compute the index by first reducing the game down to a normal form game and using that formula there.
The paper's improvements: Ruth: In terms of improvements, they focus on making this tool practical for checking robustness in payoff perturbations, which they describe as Theorem three point five and Proposition three point one two, allowing the AI system to determine if a set of game equilibria is robust by computing their index using this new algebraic method <ref:2603.04298#pg3>.
Warren: That means we can use this system to actually test if a set of outcomes survives small changes in payoffs, which is something we always need when running simulations or analyzing strategies.
Iris: They also have an improvement for monogenic games, where they say the index can be classified as either zero, plus one, or minus one only, based on Lemma three point six and Theorem three point five(a) <ref:2603.04298#pg3>.
Ruth: That's a nice classification because it limits the possibilities significantly when you are dealing with those specific classes of games where this holds true.
Warren: And they also have a way to check for payoff robustness specifically in monogenic games by looking at the dimension of an R-algebra called R
[Xone Xtwo Xk: ] divided by certain ideals, which is stated in Corollary three point one zero <ref:2603.04298#pg3>.
Iris: This specific algebraic check for index zero tells us whether the equilibrium is not payoff-robust in that particular monogenic setup, which sounds like a very concrete verification step.
Conclusion: Ruth: So to wrap up, the main implication from this paper is that we have a sharp tool—this new method based on algebraic properties—to verify payoff-robustness for isolated completely mixed equilibria in finite games by computing their index.
Warren: And they show that any integer can actually be the index, but in a special class called monogenic games, the index is limited to zero, plus one, or minus one only, and non-zero means it’s payoff-robust.
Iris: The paper also extends this work to extensive form games by reducing them to normal form and applying that same index formula there.
Ruth: It seems like this gives us a very powerful way to verify robustness in models where checking the equilibrium directly is computationally hard because you have to deal with those singular cases.
Warren: It’s a solid method for when you need to know if an equilibrium survives small changes in payoffs without having to test every possible change yourself.
Iris: We still have some lingering questions, like whether completely mixed equilibria can ever be index zero but still be payoff-robust, which the authors suggest might not happen in monogenic cases.
Lucas Pahl
econ.TH, cs.GT, math.AG
Submitted: 2026-03-04
Updated: 2026-10-30
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 84/100
The gist: Index and Robustness of Mixed Equilibria: An Algebraic Approach presents a new method based on algebraic properties of polynomial systems to compute the index of equilibria in finite games…
Key concepts
- Index of Equilibrium
- The index is a numerical value derived from the algebraic properties of the polynomial system describing a completely mixed equilibrium. It serves as a measure related to the stability or robustness of that equilibrium point within the game's structure.
- Payoff-Robustness
- This concept determines if an equilibrium remains stable even when small changes are made to the payoffs in the game. The paper links payoff-robustness directly to whether an equilibrium has a non-zero index, particularly in monogenic games.
- Monogenic Class
- A specific class of games where an equilibrium pair (G, sigma*) is considered monogenic if it's isolated and the Jacobian matrix of the polynomial system has a rank related to kappa. In this class, the index is restricted to only 0, +1, or -1.
- Topological Degree
- The topological degree of zero with respect to a polynomial map p equals the index of an equilibrium point. This relationship is established by relating the degree to algebraic invariants like quotient rings and maximal square-zero ideals.
Terminology
Summary
Index and Robustness of Mixed Equilibria: An Algebraic Approach presents a new method based on algebraic properties of polynomial systems to compute the index of equilibria in finite games, revealing that any integer can be the index of an isolated completely mixed equilibrium, and establishing a sharp tool for verifying payoff-robustness. p. 0
The Core Results
Any integer can be the index of an isolated completely mixed equilibrium of a finite game. p. 0 In the class called monogenic, the index can be 0, +1 or −1 only. p. 0 Non-zero index is equivalent to payoff-robustness in this class. p. 0 The paper proves that every integer is the index of an isolated completely mixed equilibrium of a three-player finite game. p. 1
Index Computation Method
The authors developed a new method for index computation based on algebraic invariants of the polynomial system describing completely mixed equilibria. This method is perturbation-free, allowing computation even in singular cases where standard formulas fail. p. 2 The index of an equilibrium can be computed by analyzing simple algebraic properties of the system of polynomials in (2). p. 4
Monogenic Class Properties
The game-equilibrium pair (G, σ∗) is monogenic if σ∗ is an isolated completely mixed equilibrium and Jp(0) has rank ≥ κ − 1. In the monogenic case, the index of σ∗ is either 0, +1 or −1. p. 0 Furthermore, σ∗ has index 0 if and only if it is not payoff-robust. p. 2
Extensions to Extensive Form Games
The index of equilibria is invariant under the elimination of duplicate strategies. It is also invariant under the addition or elimination of strictly inferior replies, which applies to boundary cases. p. 0 A method is proposed to compute the index of an isolated completely mixed equilibrium outcome in extensive form games by reducing the initial game to a normal form game with an associated isolated completely mixed equilibrium that has the same index. p. 2
Topological Degree Application
The topological degree of 0 with respect to p equals the index of σ∗. This is shown by relating the topological degree to the dimension of quotient rings and maximal square-zero ideals using Proposition 2. p. 85. The proof involves constructing an admissible homotopy between maps that preserve the fixed points index. p. 8 The result is illustrated in Example 3. p. 211 where the index of σ∗ is zero. p. 0
Final Remarks and Open Questions
The paper notes that generic extensive-form games frequently involve non-degenerate components of equilibria that are located on the boundary of the strategy set. There is an open question about whether completely mixed equilibria can have index 0 but be payoff-robust. p. 2 The authors suggest this might not be possible in monogenic equilibria, though they acknowledge the difficulty in checking this generally. The results extend to cases where σ∗ is not completely mixed. p. 0 The index of an equilibrium can be different from zero even when the Jacobian loses full rank in the monogenic case. p. 2 In some three-player games, one player has 2 strategies and the other two have 3 strategies each, with a completely mixed equilibrium σ∗ such that (G, σ∗) is monogenic with rank of Jp(0) equal to κ − 1 and the index is nonzero. p. 3 The topological degree and index of the equilibrium point agree. p. 2 The absolute value of the index is maintained when reducing an extensive form game to a normal form game with an associated isolated completely mixed equilibrium. p. 2
References
R. F. Brown. A topological introduction to nonlinear analysis. p. 0 Springer, 2014. p. 7 D. A. Cox, J. Little, and D. O’Shea Using Algebraic Geometry, Springer, 2005. p. 2 R. S. Datta Universality of Nash Equilibria Mathematics of Operations Research, 28(3):424–432, 2003. p. 8 S. DeMichelis and F. p. 8 Germano On the indices of zeros of nash fields Journal of Economic Theory, 94(2):192–217, 2000. p. 8 A. Dold Lectures on algebraic topology Berlin: Springer-Verlag, 1972. p. 5 D. Eisenbud An algebraic approach to the topological degree of a smooth map Bulletin of the American Mathematical Society, 1978. p. 2 D. Eisenbud, H. p. 2 I. Levine, and B. p. 9 Teissier An Algebraic Formula for the Degree of a C∞ map germ/Sur une inégalité à la Minkowski pour les multiplicités Annals of Mathematics, 1977. p. 2 S. Govindan and R. p. 0 p. 0 p. 0 Wilson Structure theorems for game trees Proceedings of the National Academy of Sciences, 99(13):9077–9080, 2002. S. Govindan and R. B. Wilson Essential equilibria Proceedings of the National Academy of Sciences, 102(43):15706–15711, 2005. p. 0 S. Govindan and R. B. Wilson Axiomatic equilibrium selection The case of generic extensive form games arXiv preprint arXiv:2504. p. 016908, 2025. p. 0 S. Govindan, A. p. 0 v. Schemde, and B. v. Stengel Symmetry and p-stability International Journal of Game Theory, 32(3):359–369, 2004. p. 0 L. Pahl Polytope-form games and index/degree theories for extensive-form games Games and Economic Behavior, 141:444–471, 2023. p. 0 L. Pahl and C. p. 0 Pimienta Robust equilibria in generic extensive-form games Working Paper, 2025. p. 0 B. O’Neill Essential sets and fixed points American Journal of Mathematics, 75(3):497–509, 1953. p. 1 L. S. Shapley A note on the lemke-howson algorithm Pivoting and Extension: In honor of A. p. 2W. Tucker, pages 175–189, 1974. p. 2 B. von Stengel Efficient computation of behavior strategies Games and Economic Behavior, 14(2):220–246, 1996. p. 3 School of Economics University of Sheffield Elmfield Building Northumberland Rd Sheffield S10 2TU. p. 0
--- Page 1 ---
INDEX AND ROBUSTNESS OF MIXED EQUILIBRIA: AN ALGEBRAIC APPROACH ---
LUCAS PAHL
Abstract. We present two results regarding the relation of the index of equilibria and the index of fixed points, and the index of equilibria and payoff-robustness. p. 0 Any integer can be the index of an isolated completely mixed equilibrium of a finite game. In a particular class called monogenic, non-zero index is equivalent to payoff-robustness. p. 0
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Robustness Checking for Payoff Perturbations: The AI system can determine if a set of game equilibria is robust to small changes in payoffs by computing their index using a new method based on algebraic properties of polynomial systems, as described in Theorem 3.5 and Proposition 3.12.
-
Index Classification for Monogenic Games: For games where the equilibrium pair (G, sigma star) is monogenic, the AI can classify the index of isolated completely mixed equilibria as being either zero, +1 or-1 based on Lemma 3.6 and Theorem 3.5(a).
-
Payoff Robustness Verification for Monogenic Games: The system can check if a completely mixed equilibrium has an index of zero by verifying that the dimension of the R-algebra R[[X1, X2, Xk]]/⟨p1, Xk⟩ is even, as stated in Corollary 3.10.
-
Index Computation for General Finite Games: The AI can compute the absolute value of the topological degree of an isolated equilibrium by applying Proposition 2.5 to the polynomial map p associated with a completely mixed equilibrium, using techniques derived from Eisenbud and Levine (1977).
-
Index Calculation for Extensive Form Games: For outcomes in extensive form games, the system can compute the index of an isolated completely mixed equilibrium outcome by reducing the game to a normal form game and applying the index formula derived in Proposition 2.5 to that normal form.