Scribd
Upload
69 views1 page

Answer Key VI

The document contains the answer key to a problem set on game theory. It discusses: 1) A mixed-strategy Nash equilibrium in an altered attack-defense game where defending the more fortified location is less likely. 2) A dominant strategy for each witness in a crime to call the police, with calling probability depending on costs and number of witnesses. 3) A correlated equilibrium for a coin flipping game between siblings that can Pareto dominate the Nash equilibria. It also discusses missing a mixed-strategy Nash equilibrium in a previous problem and provides the additional equilibrium strategies.

Uploaded by

Parmesh Gupta
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
69 views1 page

Answer Key VI

The document contains the answer key to a problem set on game theory. It discusses: 1) A mixed-strategy Nash equilibrium in an altered attack-defense game where defending the more fortified location is less likely. 2) A dominant strategy for each witness in a crime to call the police, with calling probability depending on costs and number of witnesses. 3) A correlated equilibrium for a coin flipping game between siblings that can Pareto dominate the Nash equilibria. It also discusses missing a mixed-strategy Nash equilibrium in a previous problem and provides the additional equilibrium strategies.

Uploaded by

Parmesh Gupta
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

EC 246: Decisions & Games

Answer Key to Problem Set VI

Problem 1: (i) Attacker targets “C” with probability 42.86% (targets “N” with
complementary probability); defender defends “C” with probability 33.33% (defends “N”
with complementary probability). (ii) Probability that an attack will succeed is 50%.
Comment: In the “altered” Attack-Defense game, the “more fortified location N” is attacked
with a lower probability and defended with a lower probability in equilibrium. This makes
perfect sense. If I spend a lot of effort to improve my backhand return of serve, presumably
you will serve to my backhand less often and (anticipating that) I will need to “prepare” for a
backhand return less often. Does that mean that all my effort is wasted? Of course not, if the
overall probability of my returning your serve increases. That’s exactly what happens in the
new mixed-strategy Nash equilibrium of the changed Attack-Defense game: probability that
the defense is successful in equilibrium rises to 50% irrespective of the location of attack.

Problem 2: (i) Each witness calls the police with probability {1 – [(C–Δ)/B]1/(N–1)}.
(ii) Each witness has a dominant strategy to call the police.

Problem 3: Consider a (biased) coin which comes up heads with probability p  (0, 1) and
tails with complimentary probability. The following is a correlated equilibrium: The parent
will flip the coin; if heads comes up Bob will order 2 bowls of soup and Alice will order 3
bowls of soup; while if tails comes up Bob will order 3 bowls of soup and Alice will order 2
bowls of soup. Now figure out those values of p  (0, 1) (if any) such that the specified
correlated equilibria will Pareto-dominate all the (pure and mixed) Nash equilibrium
outcomes of the of the Sibling Rivalry game.
__________________________________________________________________________
Now, regarding Question 5 in Problem Set V, I seem to be making it a habit of missing out
on some mixed-strategy Nash equilibria in simultaneous-move games. So, in the Claim
Subgame in Question 5, there is an additional mixed strategy Nash equilibrium – as pointed
out to me by Rohit Sharma – where each player mixes over {20K, 22K, 24K} with
probabilities {10/17, 4/17, 3/17}. Luckily for us, the expected payoff to each player in this
equilibrium is 19.35 < 22. Therefore, while the set of SPNE of the full game ‘expands’ to
include the following additional ‘continua of SPNE’:
[Y: {A, B; p, (1 – p) for any p  [0, 1]}, {20K, 22K, 24K; 10/17, 4/17, 3/17} if (B, B);
X: A, {20K, 22K, 24K; 10/17, 4/17, 3/17} if (B, B)];
this last set of SPNE also fail the forward induction refinement.
___________________________________________________________________________
Answers to the Sample Midterm questions:
Question 1: In the unique Nash equilibrium, Row’s strategy is {b, c; 2/3, 1/3}, and
Column’s strategy is {x, z; 4/7, 3/7}.
Question 2: H = 141, L = 120.
Question 3: In the unique Nash equilibrium, Gamma’s strategy is {stay as is, double lab
size; 7/20, 13/20}, and Delta’s strategy is {stay as is, double lab size; 13/20, 7/20}. The Nash
equilibrium expected payoff to Gamma is Rs. 130 million, and to Delta is zero.

You might also like