Course talk:CPSC522/Predicting Human Behavior in Normal-Form Games

Hi Adnan Reza,

I am glad to read your page and I learned a lot about the application of game theory in predicting human behavoir. However, I do not see your own thoughts on how successful it was and how it can be improved.

Best regards,

Ke Dai

Thank you for your suggestions. I have revised the page to include my thoughts on how it can be improved or built upon (Refer to last paragraph).

Hi Adnan,

Interesting read about improvements to Nash equilibrium.

For the comparison between models, can the formula for the log likelihood be stated? It isn't quite clear what is being compared, and how the confidence intervals are constructed.

As with the other two, I agree that the contributions of the two papers can be more explicitly stated.

For your 2-player 2-action example, I don't understand the statement "The pure-strategy Nash equilibria for this game are $(C,D)$ and $(D,C)$, where each player has no incentive to deviate." It seems that if the opponent chooses "Dare", then the player does have an incentive to deviate (chicken out). It may be better to change the example so that the Nash equilibrium is for both players to "Dare" and thus resulting in a crash. I think this would be a better lead into how "Nash equilibrium often makes counter-intuitive predictions." (Though I guess this is subjective..)

Thanks for the read,

Ricky

Hi Ricky,
Thank you for your suggestions. I have revised the page and have explicitly stated the incremental contribution of Paper 1 over Paper 2 at the end of the page. Let me try to explain what the statement you were referring to means.
Statement: The pure-strategy Nash equilibria for this game are $(C,D)$ and $(D,C)$, where each player has no incentive to deviate.
What this means is that these pairs of strategies which constitute the pure-strategy Nash equilibria are "self-enforcing", i.e. it makes each player's strategy an optimal (best) response to the other player's strategy. The main point is to show that while Nash equilibria is an intuitive notion of achieving equilibria, it is often a poor predictor of human behavior.

Hi Adnan,

Excellent page! Well-written and organized.

General comments:

• I would have liked to see more clearly what work corresponds to which papers; for example, is QLk the incremental contribution you intended to describe for the assignment, or is QCH?
• It's good that the sources for the figures are given when I click on them; but it would be good to add citations directly in the page as well.

Section-specific comments:

• Cognitive Hierarchy
  • F (j | j < m) should be enclosed in math tags, I think.
• The Modified Model (QCH)
  • • "The authors" - which authors?

Clear skies,
Jordon

Hi Jordon, Thank you for your suggestions. I have added citations to the figures directly on the page and put in the missing math tags you pointed out. I've also explicitly stated the incremental contribution of Paper 1 over Paper 2 at the end of the page.

Hi Adnan,

It is a great draft and I learned lots of things from your page. Here are some suggestions:

1. Add some motivating examples to make some definitions in Behavioral Game Theory Models part more clear and easy to understand.
2. The actual contribution of the latter paper part should described more clear.
3. Some grammar errors should be modified.

Bests,

YuYan

Hi YuYan, Thank you for your suggestions. I've revised the page and corrected all grammatical errors and/or typos. I've also explicitly stated the incremental contribution of Paper 1 over Paper 2 at the end of the page.