A definition in my book states

$ \sigma^* \in \Delta$ is a Nash equilibrium for $\Gamma$ if for all $i \in I$ and all $\mu_i \in \Delta_i, \; U_i (\sigma^*) \geq U_i(\sigma^* \backslash \mu_i)$. The set of Nash equilibria for $\Gamma$ is denoted $Eq(\Gamma)$

What does the $\backslash$ here represent, divide or element not included?

  • $\begingroup$ What do you know about the definition of the Nash equilibrium in informal words, rather than the formal mathematical definition? $\endgroup$ – 410 gone Jul 18 '16 at 16:33
  • 1
    $\begingroup$ Nothing, this is the first I've heard about it. $\endgroup$ – Sunhwa Jul 18 '16 at 16:35
  • $\begingroup$ I've skimmed the Wikipedia page, they use the notation: $$ \forall i, x_i \in S: f_i (x_i^*, x_{-1}^*) \geq f_i (x_i, x_{-1}^*) $$ Where I think the $x_{-1}^*$ stands for the nash equilibrium actions (best result by not changing) of all the other players (without player $i$). But still, I don't see how the notation in my question says this. $\endgroup$ – Sunhwa Jul 18 '16 at 16:56
  • $\begingroup$ The definition is rather odd. What's the reference? Never seen it before. In your previous comment it should probably read $-i$. $\endgroup$ – clueless Jul 18 '16 at 18:31
  • $\begingroup$ Concerning the wiki entry, you may define $x_i^* \in \arg\max_{x_i \in S_i}f_i(x_i,x_{-i}^*)$. $\endgroup$ – clueless Jul 18 '16 at 18:37

Informally, a (possibly mixed) set of strategy choices is a Nash equilibrium if no player can expect to benefit by changing strategies while the other players keep their strategies unchanged.

Here I would interpret your definition as $\Delta$ being the set of all possible strategy choices across all players, $\sigma^*$ being a set of strategy choices which will be a Nash equilibrium if it satisfies the conditions, $i$ being a player, and $\sigma^* \backslash \mu_i$ being a change in player $i$'s strategy while the other players' strategies stay the same. So I would take $\backslash$ having a type of "not included" interpretation.

  • 1
    $\begingroup$ This makes sense if we define $\sigma^* \setminus \mu_i:=(\ldots,\sigma^*_{i-1}, \mu_i,\sigma^*_{i+1},\ldots)$. However, it's more conventional and intuitive if you write $U_i(\sigma^*_i, \sigma_{-i}^*) \geq U_i(\mu_i, \sigma_{-i}^*) \forall \mu_i \in \Delta_i$, where $\sigma^*_{-i}:=(\ldots,\sigma^*_{i-1}, \sigma^*_{i+1},\ldots) \in \Delta \setminus \Delta_i$. $\endgroup$ – clueless Jul 20 '16 at 12:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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