Equivalent game forms and coalitional power (Q804482)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Equivalent game forms and coalitional power
scientific article

    Statements

    Equivalent game forms and coalitional power (English)
    0 references
    1990
    0 references
    A game form is a strategic game from which the payoffs are suppressed. It consists of a finite player set N, a finite strategy space \(S_ i\) for each player i and a map g which assigns to every n-tuple of strategies a point of an outcome space Z (or a probability measure on Z if the game form contains a non-trivial `chance player'). Introducing for each player a utility function on the outcome space Z makes the game form to a strategic game. If we identify equivalent strategies (i.e. strategies which give the same utility for every strategy (n-1)-tuple of the opponents) and delete redundant strategies (strategies which have the same payoff as a convex combination of other strategies) we find a reduced strategic game. Two game forms are called equivalent if introduction of utilities generates the `same' reduced games. Also game forms can be reduced by identifying equivalent strategies and skipping redundant strategies. Game forms are called identical if the strategy sets can be identified under a bijective map and linear dependencies between the outcome probabilities of a set of strategy n-tuples in the first game form are transformed into linear dependencies between the outcome probabilities in the second game form and vice versa. In section 2 of this paper it is proved that game forms are equivalent if and only if the reduced game forms are identical. In section 3 \(\alpha\)- and \(\beta\)-effectivity functions are associated with reduced game forms and `identity' of effectivity functions is defined. If two game forms have identical \(\alpha\)- and \(\beta\)-effectivity functions the game forms need not be equivalent but if the \(\alpha\)-effecivity function is moreover agent-additive (i.e. if \(A_ i\) is effective for \(\{\) \(i\}\), (i\(\in N)\) then \(\cap_{i\in N}A_ i\) is effective for N) then the two game forms are equivalent. Further it is proved that agent-additivity follows from superadditivity or convexity of the \(\alpha\)-effectivity function or the property of the extensive game form to have unique outcomes at all points of the game tree.
    0 references
    0 references
    game form
    0 references
    outcome space
    0 references
    reduced strategic game
    0 references
    effectivity functions
    0 references
    agent-additivity
    0 references
    superadditivity
    0 references
    convexity
    0 references
    0 references