Skip to content

Max-dominated strategy

A game-theoretic strategy that is never a best response to any admissible strategy profile of the other players, even though no single alternative need yield a strictly higher payoff against every profile.

Version
v1 · 2026-09-28 · History
Domain-specific #
10619
Domain group
Social Sciences
Origin domain
Economics & Finance
Subdomain
Game Theory → Economics & Finance

Core Idea

A max-dominated strategy is a strategy that is never a best response. Hold a player's strategy fixed and vary all admissible strategy profiles of the other players; if there is no profile at which the candidate maximizes that player's payoff, the candidate is max-dominated.

Strict domination implies max-domination, but max-domination can hold without one alternative beating the candidate everywhere. Different alternatives may outperform it in different regions of the opponents' profile space. Ties matter: a strategy tied for maximum somewhere is a best response there and is therefore not max-dominated under the usual correspondence definition.

Analysis must declare finite or continuous game, pure or mixed strategies, admissible profiles, payoff and information assumptions, and whether iterative deletion changes the game. Never-best-response elimination can simplify strategic analysis, but order, epistemic interpretation, and equivalence to dominance by mixed strategies depend on the theorem and game class actually used.

Structural Signature

Sig role-phrases:

  • player and candidate strategy. Select the decision rule being tested within a game. Constitutive object. If altered: Dominance is player-relative.
  • opponents' admissible profiles. Supply every contingency over which best-response status is checked. Constitutive domain. If altered: Pure versus mixed domains can change the result.
  • payoff function and information. Ranks the player's alternatives at each profile under the game specification. Constitutive comparison. If altered: Beliefs do not replace payoffs in the definition.
  • best-response correspondence. Records all payoff-maximizing strategies, including ties, for each profile. Identity-bearing test. If altered: Never uniquely best is weaker than never best.
  • domination witness or covering alternatives. Explains why another strategy is always at least better, possibly with different alternatives across profiles. Diagnostic structure. If altered: One global strict dominator is not required.

What It Is Not

  • Not rarely played. Observed frequency does not define best response.
  • Not never uniquely best. A tie for maximum prevents max-domination.
  • Not always strictly dominated. Different alternatives may cover different profiles.
  • Not belief-specific suboptimality. The test ranges over all admissible opponent profiles.

Scope of Application

The concept is used in game theory, strategic-form simplification, rationalizability, mechanism analysis, learning models, dominance solvability, computational game solving, and economic pedagogy.

  • Normal-form games. Checks payoff tables.
  • Mixed strategies. May expand profile and dominance sets.
  • Rationalizability. Relates to never-best-response deletion.
  • Algorithms. Prunes strategy spaces cautiously.
  • Mechanism analysis. Identifies choices unsupported by any contingency.

Clarity

Report player, game and information structure, candidate strategy, opponents and their admissible pure/mixed profile space, complete payoff function, tie and best-response convention, calculation or optimization method, witness regions/covering alternatives, whether one strict dominator exists, iteration round and deletion order, theorem relating deletion to rationalizability, and numerical tolerances.

Manages Complexity

The test compresses an entire opponents' profile space into a supportability verdict while distinguishing a locally changing cover of better replies from one globally dominating alternative.

Abstract Reasoning

  1. Fix player, strategy sets, payoff, information, and pure/mixed domain.
  2. Compute the candidate's payoff over every admissible opponents' profile.
  3. Compute the full best-response correspondence including ties.
  4. Certify absence or exhibit a supporting profile.
  5. If deleting iteratively, restate the reduced game and theorem conditions.

Knowledge Transfer

The never-optimal-under-any-environment pattern transfers to decision pruning, but payoffs, admissible environments, ties, and strategic feedback must be reconstructed rather than copied.

Examples

Canonical

In a finite payoff table, the candidate strategy is below some alternative in each opponent column, although the better alternative changes by column; because it is never tied for column maximum, it is max-dominated but need not be strictly dominated by one pure strategy.

Mapped back: player and candidate strategy → tested row strategy; opponents' admissible profiles → all columns; payoff function and information → declared row payoffs; best-response correspondence → column maxima including ties; domination witness or covering alternatives → different higher rows across columns.

Applied / In Practice

A solver partitions a mixed-opponent simplex into best-response regions, verifies that one strategy has an empty region within numerical tolerances, and records whether a mixed strict-dominance certificate also exists.

Mapped back: player and candidate strategy → solver target; opponents' admissible profiles → mixed-strategy simplex; payoff function and information → expected-payoff model; best-response correspondence → polyhedral maximum regions; domination witness or covering alternatives → empty region and optional certificate.

Structural Tensions

T1: pruning vs. strategic fidelity. Deletion simplifies a game while changed strategy sets can alter later reasoning. Diagnostic: Which deletion theorem and order apply?

T2: pure simplicity vs. mixed completeness. Pure profiles are easy to inspect while mixed profiles may change supportability. Diagnostic: Which admissible domain defines the claim?

T3: exact ties vs. numerical tolerance. Best-response membership is exact while computation uses finite precision. Diagnostic: How were near-ties certified?

Structural–Framed Character

Max-domination is structural. Candidate, environment profiles, payoff ordering, and best-response correspondence fully define it; interpretation does not change the mathematical identity. Evaluative weight and human-practice dependence are low; origin is game theory; vocabulary travels with optimization semantics; transfer recognizes a never-optimal structure. Its portable skeleton is Unsupported Option, a prospective future-prime candidate. Its character: an available strategy with no environment in which it reaches the payoff maximum.

Structural Core vs. Domain Accent

Skeletal core. Test whether any admissible environment supports an option as optimal.

Domain-bound accent. Players, opponent profiles, payoff functions, mixed strategies, ties, and best responses define max-domination.

Why not prime. Unsupported-option reasoning travels; this is a game-theoretic strategy relation.

  • Dominance. Stronger neighboring comparison requiring exact live-signature review.
  • Best response. The correspondence whose empty support defines the concept.

Neighborhood in Abstraction Space

Max-dominated strategy sits in a crowded region of the domain-specific corpus (19th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Strategic Games & Equilibrium Concepts (13 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08

Not to Be Confused With

  • Strictly dominated strategy. Tell: One global superior alternative or changing cover?
  • Never uniquely best. Tell: Does a tie make it a best response?
  • Low-frequency strategy. Tell: Behavioral rarity or payoff structure?
  • Belief-dominated choice. Tell: One belief or every admissible profile?

References

  • Frozen Wikipedia discovery revision: https://en.wikipedia.org/wiki/Max-dominated_strategy (revision 1292170039).
  • Preserved source candidate: http://www.springerlink.com
  • Preserved source candidate: https://web.archive.org/web/20030417141513/http://www.springerlink.com/
  • Preserved source candidate: https://doi.org/10.1007%2F978-3-540-92185-1_59

The frozen Wikipedia revision is discovery provenance. The retained source set was reviewed for identity, formal or operational relation, and scope. The encyclopedia's structural synthesis is bounded to those claims; a thin authority surface is recorded as a nonblocking source-strengthening repair rather than concealed.