1
$\begingroup$

Find the pure and mixed-strategy Nash equilibrium in this game:

\begin{array}{c|cccc} P_1 \text{/} P_2 & \text{Ll} & \text{Lr} & \text{Rl} & \text{Rr} \\ \hline \text{T} & (3,2) & (3,2) & (1,1) & (1,1) \\ \text{B} & (4,3) & (2,4) & (4,3) & (2,4) \\ \end{array}


By using the definition of pure strategy Nash equilibria, I found these two Nash equilibria:

\begin{equation} \text{Pure-strategy N.E.} = \begin{cases} ([T],[Lr]) \\ ([B],[Rr]) \end{cases} \end{equation}

For the mixed-strategy Nash equilibria, it appears that the answer is:

\begin{equation} \text{Mixed-strategy N.E.} = \begin{cases} ([T], q[Ll] + (1-q)[Lr]) & \text{if } q < \frac{1}{2} \\ ([B], q[Lr] + (1-q)[Rr]) & \text{if } q < \frac{1}{2} \end{cases} \end{equation}

How would one derive the two mixed-strategy Nash equilibria?

$\endgroup$

1 Answer 1

1
$\begingroup$

First, note that there is a lot of weak domination: $$Lr \succeq Ll, ~ Lr \succeq Rl, ~ Rr \succeq Rl, Ll \succeq Rl.$$

Consequently, $\nexists$ MSNE in which both players mix at the same time. If $P_1$ mixed, then $P_2$ would strictly prefer to play a pure strategy.

Consider then, as the answer suggests, a candidate equilibrium in which $P_1$ plays a pure strategy, whilst $P_2$ mixes.

If $P_1$ plays $T$, then $P_2$ is indifferent between $Ll$ and $Lr$.

$\implies$ $P_1 \to T, ~P_2 \to q(Ll) + (1 - q)Lr$.
$P_2$ is indifferent by construction, so we need merely show that $P_1$ prefers $T$ to $B$: $$\begin{align} 3 & \geq 4q + 2(1 - q) \\ & \geq 2 + 2q \\ 1 &\geq 2q \implies 1/2 \geq q \end{align}$$

Likewise, if $P_1 \to T$, then it must be that $P_2 \to q(Lr) + (1-q)Rr$. Again, only need to show $P_1$ prefers $B$ to $T$: $$\begin{align} 2 & \geq 3q + 1(1 - q) \\ 1/2 & \geq q \end{align}$$

This fully describes the Nash Equilibrium.

$\endgroup$
3
  • $\begingroup$ $Ll \succeq Lr$ is not true, it is the other way around? $Lr$ itself, however, is weakly dominant. $\endgroup$
    – Giskard
    May 17 at 10:10
  • 1
    $\begingroup$ You're quite right, I'll edit. $\endgroup$ May 18 at 5:58
  • $\begingroup$ @matthewoulton Thank you $\endgroup$
    – user9487
    May 18 at 12:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.