Game Theory
By- Dr. B. J. Mohite
9850098225
Meaning
Game Theory is mathematical model which is concerned
with the study of decision making in situation where two
or more opponents are involved under condition of
conflicting interests or goals
Developed in1950s by mathematicians John von
Neumann and economist Oskar Morgenstern Designed to
evaluate situations where individuals and organizations
can have conflicting objectives
Game Theory is a set of tools and techniques for
decisions under uncertainty involving two or more
intelligent opponents in which each opponent aspires to
optimize his own decision at the expense of the other
opponents.
In game theory
an opponent is referred to as player.
Each player has a number of choices, finite or infinite,
called strategies.
The outcomes or payoffs of a game are summarized
as functions of the different strategies for each player
B Player
B1 B2 B3
Strategies
A1 2 -8 4
Payoffs
A
A2 7 10 6
Payoff Matrix
Profit to A Loss to Profit to B Loss to
B A
Essential Features of Game
Theory
A competitive situation is called a game if it has
following features:
Finite number of competitors.
Finite number of action.
Knowledge of alternatives, Choice & Outcome or
Gain.
Choice of opponent.
Terms used:
Two Person Zero-sum game : It is the situation
which involves two persons or players and gains
made by one person is equal to the loss incurred
by the other.
n-persons game : A game involving n persons is
called a n-person game.
Payoffs : Outcome of a game due to adopting the
different courses of actions by competing players
in the form of gains or losses for each of the
players is known as payoffs.
Classification of Games
Two-Person Game – A game with 2 number of players.
Zero-Sum Game – A game in which sum of amounts
won by all winners is equal to sum of amounts lost
by all losers.
Non-Zero Sum Game – A game in which the sum of
gains and losses are not equal.
Pure-Strategy Game – A game in which player selects
only one strategy & ignores other strategies to
Maximize gain or Minimize loss.
Mixed-Strategy Game – A game in which player
selects more than one strategy to Maximize gain or
Payoff matrix – In a game, the gains or losses, resulting
from different moves & counter moves, when represented
in the form of a matrix are known as pay off matrix.
Positive Payoff- Profit to Player A, Loss to Player B
Negative Payoff- Loss to Player A & Profit to Player B
Maximin Criteria/Principle – The maximizing player
lists his minimum gains from each strategy and selects
the strategy which gives the maximum out of these
minimum gains.
Minmax Criteria/Principle – The minimizing player lists
his maximum loss from strategy and selects the strategy
which gives him the minimum loss out of these maximum
losses.
Value of Game – The maximum guaranteed gain to the
Saddle Point
Saddle point is the equilibrium point for both player.
(Maximin=Minimax)
It is the smallest value in its row & largest value in its
column.
Steps to find out Saddle Point:-
1. Select the minimum value of each row & put a circle O around maximum value.
2. Select the maximum value of each column & put square □ around minimum
value. B
Row
3. The value with both circle
B1 andB2 square Minimum
B3is the saddle point.
A1 2 8 4 2
A
A2 7 10 6 6 Maximin
Column
7 20 6 Value of
Maximum
game
Minimax
Strategy
It is the pre-determined rule or action by which each player
decides his course of action from list available to him.
It is course of action taken by one of the participants in a game.
Types of Strategy
Pure Strategy- Only One strategy has probability value 1 & other 0
P1=0, P2=1, p3=0
Mixed Strategy- Sum of probabilities of strategies=1 & individual
strategy probability value less than 1
P1=0.33, P2= 0.26, P3=0, P4=0.41
Games with Pure Strategy
Find the optimum strategies of the player in the following game
B
Row
B1 B2 B3 Minimum
A1 2 8 4 2
A A2 7 10 6 6 Maximin
A3 3 5 4 3
Column
7 20 6
Maximum
Minimax
Here, Maximin=Minimax, Hence game has Saddle point at the cell
intersecting A2 & B3 strategies. & Value of the game (V)= 6
A[P1, P2,P3]= A(0, 1, 0)
B[P1, P2, P3]= B(0, 0, 1)
Hence Player A must select Strategy A2 to maximize profit &
Games with Mixed Strategies
Find the optimum strategies of the player in the following game
B
Row
B1 B2 Minimum
A1 9 7 7 Maximin
A
A2 5 11 5
Column
9 11
Maximum
minimax
Here, Maximin ≠ Minimax, Hence game has No Saddle point, so we
can solve using Mixed strategy.
B
Oddment
Probability
B1 B2 s
6/
A1 9 7 (11-5)=6 (6+2)=6/8=3/4
A
2/
A2 5 11 (9-7)= 2 (6+2)=2/8=1/4
Oddments (11-7)= 4 (9-5)=4
4/(4+4)=4/8=1/2 4/(4+4)=4/8=1/2
A(P1, P2)=A(3/4, 1/4)
B(P1, P2)= B(1/2, ½)
Value of Game (V)=) = (54+10)/8= 64/8 =8
Dominance Method
The Principle of Dominance states that, if the strategy of a player
dominates over the other strategy in all condition then the later
strategy is ignored.
Rules to be followed :
If all the elements of a row (say Xth row) of a pay off matrix are
less than or equal to (≤) the corresponding each element of
the other row (say Yth row) then the player A will never choose
the Xth strategy OR Xth row is dominated by Yth row. Then
delete Xth row (least values).
If all the elements of a column (say Xth column) are greater than
or equal to the corresponding elements of any other column (say
Yth column) then Xth column is dominated by Yth column.
Then delete Xth column. (Highest values)
If ith Row values are ≤ average of remaining row values, delete ith row
If ith Column values are ≥ average of remaining column values, delete ith row
The following table represent the payoff matrix with
respect to player A. Solve it optimally using dominance
property with Pure strategy
B1 B2 B3 B4 B5
A1 4 6 5 10 6
A2 7 8 5 9 10
A
A3 8 9 11 10 9
A4 6 4 10 6 4
B
Row
Minimu
B1 B2 B3 B4 B5 m
A1 4 6 5 10 6 4
A2 7 8 5 9 10 5
A Maximi
A3 8 9 11 10 9 8
n
A4 6 4 10 6 4 4
Col
8 9 11 10 10
Maximum
Game has Maximin
Minimax = Minimax,
Hence saddle point at strategy A3 & B1
Player A(P1, P2, P3, P4)= A(0, 0, 1, 0)
Player B(P1, P2, P3, P4, P5)= B(1, 0, 0, 0, 0)
Value of game (V)=8
B
Total
B1 B2 B3 B4 B5
A1 4 6 5 10 6 31
A2 7 8 5 9 10 39
A A3 8 9 11 10 9 47
A4≤A3,
A4 6 4 10 6 4 30 Delete
A4
B
Total
B1 B2 B3 B4 B5
A1≤A3,
A1 4 6 5 10 6 31 Delete
A A1
A2 7 8 5 9 10 39
A3 8 9 11 10 9 47
B
Total
B1 B2 B3 B4 B5
A2 7 8 5 9 10 39 No dominance
A
A3 8 9 11 10 9 47 No dominance
B
B1 B2 B3 B4 B5
A2 7 8 5 9 10
A
A3 8 9 11 10 9
Total 15 17 16 19 19
B4≥B1, Delete B4
B
B1 B2 B3 B5
A2 7 8 5 10
A
A3 8 9 11 9
Total 15 17 16 19
B5≥B1, Delete B5
B
B1 B2 B3
A2 7 8 5
A
A3 8 9 11
Total 15 17 16
B2≥B1, Delete
B
B1 B3
A2 7 5
A
A3 8 11
15 16
Total
No Dominance No Dominance
B B
Total
B1 B3 B1 B3
A2≤A3, A A3 8 11
A2 7 5 12 Delete
A A2 B3≥B1, Delete
B3
A3 8 11 19
B
B1 Player A(P1, P2, P3, P4)= A(0, 0, 1, 0)
A A3 8 Player B(P1, P2, P3, P4, P5)= B(1, 0, 0,
0, 0)
Value of game (V)=8
The following table represent the payoff matrix with
respect to player A. Solve it optimally using dominance
property with Mixed strategy
B1 B2 B3 B4
A1 5 -3 3 4
A A2 -4 5 4 5
A3 4 -4 -3 3
B
Minimu
B1 B2 B3 B4
m
A1 5 -3 3 4 -3 MaxiMin
A A2 -4 5 4 5 -4
A3 4 -4 -3 3 -4
Maximum 5 5 4 5
MiniMax
Game has Maximin ≠ Minimax, Hence No saddle point
B
B1 B2 B3 B4 Total
A1 5 -3 3 4 9
A A2 -4 5 4 5 10
A3 4 -4 -3 3 0 A3≤A1, Delete A4
B
B1 B2 B3 B4 Total
A1 5 -3 3 4 9 No dominance
A
A2 -4 5 4 5 10 No dominance
B
B1 B2 B3 B4
A1 5 -3 3 4
A
A2 -4 5 4 5
Total 1 2 7 9
B4≥B2, Delete B4
B
B1 B2 B3
A1 5 -3 3
A
A2 -4 5 4
Total 1 2 7
No Dominance No Dominance No Dominance
B
Avg B1 &
B3
Compare B2
Average of B1 & A
A1 1 3
B2 with B3 A2 1/2 4
B
Row
Probability
B1 B2 Oddments
9/
A1 5 -3 5-(-4)= 9
(9+8)=9/17
A
8/
A2 -4 5 5-(-3)=8
(9+8)=8/17
Col. Oddments 5-(-3)=8 5-(-4)=9
A(P1, P2, P3)=A(9/17, 8/17,9/ 0)
Probability 8/(8+9)=8/17 (8+9)=9/1
B(P1, P2, P3, P4)= B(8/17, 9/17, 7 0, 0)
Value of Game (V)=)
= (45 - 32)/17
= 13/17
= 0.76
Game theory using Matrix
method B
B
B2 B3
1
A1 7 1 7
A A2 9 -1 1
A3 5 7 6
To check Saddle point
above B
Minimu
B m
B2 B3
1
A1 7 1 7 1
Left
A2 9 -1 1 -1
A
Maximi
A3 5 7 6 5
n
Maximu
9 7 7
m
Here, Maximin ≠ Minimax, Hence game
MiniMa
has No Saddle point
x
Dominance B
method Total
B1 B2 B3
No
A1 7 1 7 15 Dominance
No
A A2 9 -1 1 9 Dominance
No
A3 5 7 6 18 Dominance
Matrix Total 21 7 14
method
No
No Dominance
B No Dominance
Dominance
B1-
B2-B3 Oddments
B1 B2 B3 B2
A1 7 1 7 6 -6 6
A A2 9 -1 1 10 -2
A3 5 7 6 2 1
A1-A2 -2 2 6
A2-A3 4 -8 -5
Oddments
Oddments for strategy A1
using determinants
B
B1- Oddmen
B B2-B3
B2 B3 B2 t
1
A1 7 1 7 6 -6 6
A A2 9 -1 1 10 -2 6
A3 5 7 6 2 1 48
A1-A2 -2 2 6
A2-A3 4 -8 -5
Oddmen 3 1
t
8
8 4
B
B1- B2- Oddme
Probability
B1 B2 B3 B2 B3 nt
6/
A1 7 1 7 6 -6 6 (6+6+48)=6/60=1/1
0
6/
A A2 9 -1 1 10 -2 6 (6+6+48)=6/60=1/1
0
48/
A3 5 7 6 2 1 48 (6+6+48)=48/60=8/
10
A(P1, -2P2, P3)=A(1/10,
A1-A2 2 6
1/10, 8/10)
A2-A3 4 -8 -5
B(P1, 38
Oddment P2, P3)=
14 19/30,
8 7/30, 2/15)
Value
38/60
of Game
=19/3
14/60 (V)=)
8/60
= 7/30 = 2/15
0
= 336/60 = 5.60
Game theory using Equal game method/ Algebraic
method
Solve following game problem using equal game
method B
B
B2
1
A1 4 1
A
A2 2 3
Calculate value of P1
B
Probability (4*P1)+(2*[1-
B1 B2 P1])=(1*P1)+(3*[1-P1])
A1 4 1 (P1) 1/4 4P1 + 2 – 2P1= P1 + 3 -3P1
A
A2 2 3 (1-P1) 3/4
4P1-2P1- P1 + 3P1 = 3 – 2
Q1 (1-Q1)
Probability 4P1 = 1
P1= 1/4
B Calculate value of Q1
Probability
B1 B2 (4*Q1)+(1*[1-Q1])=(2* Q1)+ (3 * [1-
A1 4 1 (P1) 1/4 Q1])
A
A2 2 3 (1-P1) 3/4 4Q1 + 1 –Q1 =2Q1 +3 -3Q1
Q1 (1-Q1) 4Q1 –Q1 -2Q1 + 3Q1 = 3-1
Probability
1/2 1/2 4Q1= 2
Q1= 1/2
Value of game (V)
= (4 * 1/4 ) + (2* 3 /4)
= 4/4 + 6/4
= 1 + 3/2
= 5/2
Player A(P1, P2)= A(1/4, 3/4)
Player B(P1, P2)= B(1/2,1/2)
Value of game (V)=5/2
Game theory using Algebraic method (3X3 matrix)
Solve following game problem
B
B1 B2 B3
A1 1 -1 -1 Equations /Relations for Player ‘A’
A A2 -1 -1 3 X1- X2 - X3 =V -------------------------(1)
A3 -1 2 -1 -X1 -X2 + 2X3=V -----------------------(2)
-X1 + 3X2 - X3=V --------------------- (3)
B
X1 + X2 + X3 =1 ------------------------ (4)
B
B1 B3 Equations /Relations for Player ‘B’
2
A1 1 -1 -1 X1 Y1 - Y2 - Y3=V --------------------- (5)
A A2 -1 -1 3 X2 -Y1- Y2 + 3Y3=V --------------------- (6)
A3 -1 2 -1 X3 -Y1 + 2Y2 - Y3=V --------------------- (7)
Y1 Y2 Y3 Y1 + Y2 + Y3=1 --------------------- (8)
To calculate value of X1 Add eq. 1 Equations /Relations for Player
‘A’
&4 X1- X2 - X3 =V -------------------------
X1- X2 - X3 =V -------------------------(1) (1)
X1 + X2 + X3 =1 ----------------------- (4) -X1 -X2 + 2X3=V ----------------------
(2)
2X1 =V+1 Therefor X1= (V+1) / 2 -X1 + 3X2 - X3=V --------------------
(3)
To calculate value of X3 Add eq. 2 X1 + X2 + X3 =1 -----------------------
(4)
&4 Equations /Relations for Player
-X1 -X2 + 2X3=V ---------------------- (2) ‘B’
X1 + X2 + X3 =1 ----------------------- (4) Y1 - Y2 - Y3=V ---------------------
(5)
3X3 =V+1 Therefor X3= (V+1)/3 -Y1- Y2 + 3Y3=V ---------------------
(6)
-Y1 + 2Y2 - Y3=V ---------------------
(7)
To calculate value of X2 Add eq. 3 Y1 + Y2 + Y3=1 ---------------------
&4 (8)
-X1 + 3X2 - X3=V -------------------- (3)
X1 + X2 + X3 =1 ----------------------- (4)
4X2 =V+1
To calculate value of Y1 Add eq. 5 Equations /Relations for Player
‘A’
&8 X1- X2 - X3 =V -------------------------
Y1 - Y2 - Y3=V --------------------- (5) (1)
Y1 + Y2 + Y3=1 --------------------- (8) -X1 -X2 + 2X3=V ----------------------
(2)
2Y1=V+1 Therefore Y1= (V+1)/2 -X1 + 3X2 - X3=V --------------------
(3)
To calculate value of Y3 Add eq. 6 X1 + X2 + X3 =1 -----------------------
(4)
&8 Equations /Relations for Player
-Y1- Y2 + 3Y3=V --------------------- (6) ‘B’
Y1 + Y2 + Y3=1 --------------------- (8) Y1 - Y2 - Y3=V ---------------------
(5)
4Y3=(V+1) Therefore Y3 = (V+1)/ 4 -Y1- Y2 + 3Y3=V ---------------------
(6)
-Y1 + 2Y2 - Y3=V ---------------------
(7)
To calculate value of Y2 Add eq. 7 Y1 + Y2 + Y3=1 ---------------------
&8 (8)
-Y1 + 2Y2 - Y3=V --------------------- (7)
Y1 + Y2 + Y3=1 --------------------- (8)
3Y2=(V+1) Therefore Y2= (V+1) /3
Put values of X1, X2 and X3 in equation 4Equations /Relations for Player
‘A’
+ X1- X2 - X3 =V -------------------------
(1)
-X1 -X2 + 2X3=V ----------------------
(2)
-X1 + 3X2 - X3=V --------------------
(3)
X1 + X2 + X3 =1 -----------------------
V= (12-13)/13 = -1/13 (4)
Equations /Relations for Player
X1= (V+1)/2 = (-1/13 + 1)/2 = (12/13)/2‘B’
Y1 - Y2 - Y3=V ---------------------
=6/13 (5)
X2= (V+1)/4 = (-1/13 +1)/ 4 = (12/13)/4 -Y1- Y2Optimum
+ 3Y3=V ---------------------
Strategy for
=3/13 (6)
Player ‘A’ =( 6/13, 3/13,
-Y1 + 2Y2 - Y3=V ---------------------
X3= (V+1)/3 = (-1/13 +1)/ 3 = (12/13)/3 (7) 4/13)
=4/13 Y1 + Y2 + Y3=1 ---------------------
(8) Optimum Strategy for
Player ‘B’ =( 6/13, 4/13,
Y1= (V+1)/2 = (-1/13 + 1)/2 = (12/13)/2 3/13)
=6/13
Y2= (V+1)/3 = (-1/13 +1)/ 3 = (12/13)/3 Value of Game (V)= -
Limitations of Game Theory
Infinite number of strategy
Knowledge about strategy
Zero outcomes
Risk & uncertainty
Finite number of competitors
Certainty of Pay offs
Rules of game.
Thank You!...
For any queries,
Contact: Dr. B. J. Mohite
9850098225
