Proposition 3.2 Consider a two-player game G = (Am⇥n, Bm⇥n), a subset S0 of the row player’s pure
strategies, and a distinguished strategy i for the row
player. We can determine in polynomial time (in the
size of the game) whether there exists a mixed strategy
x, that places positive probability only on strategies in
S0 and dominates the pure strategy i. Similarly, for the
column player, a subset S0 of the column player’s pure
strategies, and a distinguished strategy j for the column
player. We can determine in polynomial time (in the size
of the game) whether there exists a mixed strategy y, that
places positive probability only on strategies in S0 and
dominates the pure strategy j. This applies both for strict
and weak dominance (Conitzer & Sandholm 2005).