2
$\begingroup$

In the beginning of the mahjong anime/manga Akagi, the titular character has just come in from what is called a game of chicken, which is described as follows:

Two cars head for a cliff at full speed. It's a game of reckless disregard for life where the person who first brakes loses.

No conditions for ties were given in the anime, but I will assume that:

  • Driving over the cliff is the least preferable outcome regardless of what happens to the other person

  • Losing is preferable to driving over the cliff

  • Being in a tie is preferred over losing, and winning is the most preferred outcome

I believe that the games aren't equivalent:

  • Unlike the description of the game of chicken given on Wikipedia, there is a weakly dominant strategy in the game in Akagi (i.e. braking at the edge of the cliff). But I'm not sure if the existence of weakly dominant strategies are preserved under isomorphisms of games.

  • The Akagi version of chicken has a pure Nash equilibrium where both players take the same action (and both obtain an outcome that isn't their most preferred outcome). This outcome (in terms of preference rankings) doesn't exist in the standard formulation of the game of chicken. However, thinking about it, it might be better to view the Akagi case as something continuous that might then permit mixed strategies (with regards to when the player chooses to brake).

Is my reasoning here correct, or is there a better way to view this?

Update: I should add that in the Akagi version, the choice is when to brake, which lies on a continuum in some interval of $\mathbb{R}$. (I'm not entirely sure how to formulate this mathematically, since we also need to allow for the possibility that the player drives off the cliff, but this is beside the point.) However, as previously stated, braking right at the edge should be weakly dominant, at least from an intuitive perspective.

Conversely, in the standard version, there is a binary choice ("swerve" or "straight") at a single point in time. Unless it's possible to rule out all but a few cases in the Akagi version, it would seem that the games can't possibly be equivalent, since we can't "reduce" the Akagi case to the normal game of chicken. (There should be no isomorphism between them if the cardinalities of the set of choices are different.) But that's exactly the thing that I'm not completely sure about.

Moreover, I'm not sure if I'm analysing the equilibria of the Akagi case properly, particularly since it strikes me as a scenario where mixing might do better (but I'm not familiar with mixed strategies in continuous cases, and I'm too busy at the moment to test somewhat "simpler" cases).

$\endgroup$
  • $\begingroup$ If you could express your game in the form of payoff matrices it will be a lot more understandable and will allow us to answer the question. $\endgroup$ – R B Feb 9 '15 at 9:48
  • $\begingroup$ @RB: I'll see if I can do that; I'm not entirely sure if there's a simple way to do that. $\endgroup$ – user169 Feb 9 '15 at 9:53
3
$\begingroup$

As you stated the strategy set in the traditional chicken game has two elements and in the Akagi version it is an interval, so there will be no bijective mapping between the two. If I misunderstood and you are using a different notion of equivalence please clarify.

Also as you state in the Akagi version there are weakly dominant strategies but in the traditional version there are none. I believe the story of the traditional version is rather about two cars driving towards each other and you have to decide whether to swerve at the last moment or trust that the other driver will. Even if you were to extend the strategies to allow swerving before the last moment this game would differ from the Akagi version, because here not swerving at all is better than swerving at the last moment if the other driver swerves at last moment whereas swerving at the last moment is better if the other does not swerve at all, hence there are no weakly dominant strategies.

$\endgroup$
  • $\begingroup$ That's more or less what I was thinking of. The main thing that I wasn't entirely sure of was whether or not by eliminating dominated strategies in the Akagi case, we might have a result similar to the standard game. But I think my analysis of (one of) the equilibriums in the Akagi case still covers that. $\endgroup$ – user169 Feb 11 '15 at 0:38

Your Answer

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