Show that in any 2 player symmetric game there is always a symmetric Nash equilibrium where both players play the same strategy
I read the proof for n-player game in page 12 here and all the permutations made it really hard for me to read and understand..Does anyone have a proof or can help me simplify it for n=2?
Thanks
In a non-degenerate game there is an indirect way through an argument provided by the Lemke Howson algorithm, namely that any non-degenerate bimatrix game (here I added finite to your assumptions, otherwise this seemingly obvious statement is not true: there are infinte/discontinuous symmetric games with only asymmetric Nash equilibria) has an odd number of equilibria. Hence, knowing that
you can conclude that there is an even number of equilibria of the form $(x,y)$ with $x\neq y$ which imiplies that there must a $x^S$ such that $(x^S,x^S)$ is an equilibrium.
Proof of point 2. Let $\alpha(x,y)$ denote the payoff function of player I, where $(x,y)\in X\times X$, where $X$ is the common strategy space. Let $\beta(x,y)$ denote the payoff function of player II. By symmetry $$\beta(x,y)=\alpha(y,x)\tag{sym}$$ Now, assume that $(x',y')$ is a Nash equilibrium of this bimatrix game, then $$\alpha(x',y')\ge\alpha(x,y'), \quad \forall x\in X\tag1$$ and $$\beta(x',y')\ge \beta(x',y), \quad \forall y \in X\tag2$$ Rewrite the conditions in view of the symmetric property $(\text{sym})$to obtain that $$\alpha(y',x')\ge \alpha(y,x'),\quad \forall y\in X\tag{2'}$$ and $$\beta(y',x')\ge \beta(y',x),\quad \forall x\in X\tag{1'}$$ Now, $(1')$ and $(2')$ together imply that $(y',x')$ is a Nash equilibrium of the simultaneous move game.