# Signaling in zero-sum games?

The foundation of this question is a bit vague (based on something overheard at a conference) but I'm hoping someone here can provide some clarification.

I overheard a conversation between two people at a conference saying that signaling doesn't really make sense in zero-sum games. Unfortunately, I don't have much more context than that.

I have continued to think about this statement and really can't figure out why it doesn't really make sense.'' The zero-sum aspect only talks about the payoffs, where the signaling aspect is concerned with how information is revealed as a function of the players' actions. I don't see why a game being zero-sum should preclude signaling.

Question: Can signaling occur in zero-sum games? Please provide a link to a paper/example if so.

• This is an interesting question, but it would be probably benefit from being clarified. I am guessing that what you mean is: Given a simultaneous game $G$ with strategy sets $S_1$,$S_2$ there is a set of equilibria $E$. If we consider the game variant $G'$ where player $1$ is allowed to send an observable signal (cheap talk or only zero sum?) the set of equilbria $E'$ is such that you can map each element $E'$ to an element of $E$ with the same elements of $S_1$ and $S_2$ being chosen. – Giskard Sep 29 '19 at 7:31
• (Feel free to use any part of this should you choose to edit the question to clarify.) – Giskard Sep 29 '19 at 7:32
• This calls to mind Crawford and Sobel's cheap talk model. If the sender and receiver's interests are diametrically opposed, then there is no equilibrium in which information is transmitted. – afreelunch Oct 1 '19 at 11:24