Saddle Point:
In a two-person zero-sum game (X,Y,A), a pair of actions (x, y) is a saddle point if
A(x, y) = max x' X A(x', y) = min y' Y A(x, y'). (Duersch, Oechssler, & Schipper, 2010)
In other words, an outcome in a matrix game is called a Saddle Point if the entry at that
outcome is both less than or equal to any in its row, and greater than or equal to any entry in
its column.
Ilustration of saddle point:
As per the conditions for saddle point stated above, the saddle point for the illustrative
example is (x3,y1). At (x3,y1), A is maximum compared to (x1,y1) and (x2,y1) which
satisfies A(x, y) = maxx' X A(x', y). Again at (x3,y1) A is minimum compared to (x3,y2) and
(x3,y3) which satisfies A(x, y) = miny' Y A(x, y').
A Two-Person, Zero-Sum Game may involve one of the following cases:
(i) single saddle point
(ii) more than one saddle points
(iii) no saddle point.
Saddle point can be obtained easily with the help of Minimax Theorem stated by John von
Neumann.
Since, Maximin value = Minimax value, the game has a saddle point. Thus, it can be solved
by pure strategy. Thus optimal strategy for Player 1 is A2 and for Player 2 is B3. The value of
the game is 1 and the game is not fair since a game that has a value of 0 is said to be a fair
game.
7.2.2 Two-Person, Zero-Sum Game with two saddle points:
Since, there are two Maximin values being equal to one Minimax value, the game has two
saddle points, at A1-B1 and A3-B3 positions. Thus optimal strategy for Player 1 is either A1
or A3 and for Player 2 is B1. The value of the game is 3 and the game is not fair since a game
that has a value of 0 is said to be a fair game.
In addition, it can be illustrated that "The Minimax Criterion" would yield the same result.
Since, Maximin value Minimax value, the game has no saddle point. Thus, it can't be
solved by pure strategy, rather it need to be solved by mixed strategy.
A mixed strategy problem can be solved by Graphical method (El-Kareem, 2010). But, for
that purpose, initially we need to apply Dominance property to eliminate any Dominated
Strategy if there is any.
For each of the pure strategies available to Player 2, the expected payoff for player 1 will be :
(El-Kareem, 2010)
7.4.1 Graphical Solution to obtain Mixed Strategy Probabilities and Expected Payoff:
12
From graph: x2 = 17
12
Therefore, x1 = 1 17 = 17
12
50
12
50
Expected value : 10 10 ( 17 ) = 17
2 + 7 ( 17 ) =
17
50
0 1 + 5 y2 = 17
10
we get, y1 = 17 ; y2 = 17
12
Thus, the optimal mixed strategy for Player 1 (x1, x2, x3) = ( 17 , 17 , , 0 ) and
7
10