Hi Chalie,

If you could just e-mail me the code, that'd be great.

Thanks again for participating!

 - Kevin

On Fri, Jul 4, 2014 at 1:26 PM, Evans, Charles <charles.evans.13@ucl.ac.uk> wrote:

#############################################################
# Please fill out the following information about your team 
#############################################################

Team Name: KuhnYouHandleIt
Agent Name (can be the same as team name): KuhnYouHandleIt


For each team member, please list the following:

Charlie Evans - Team leader, charles.evans.13@ucl.ac.uk, graduate student at UCL.

Was this submission part of an academic class project?  What level of class
(undergraduate/graduate)? 

Part of my thesis at graduate level.


###########################################################################
# Please provide as much information about your agent as possible as the
# competition organizers are very interested in knowing more about the
# techniques used by our competitors.
###########################################################################

1) Is your agent dynamic?  That is, does its strategy change throughout the
course of a match, or is the strategy played the same throughout the match? No

2) Does your agent use a (approximate) Nash equilibrium strategy? No

3) Does your agent attempt to model your opponents?  If so, does it do so
online during the competition or offline from data (e.g., using logs of play or
the benchmark server)? 

Counts the number of times an opponent performs a certain action at a decision point in the game tree. Calculates expected values from those counts.

4) Does your agent use techniques that would benefit from additional CPU time
during the competition? No

5) Does your agent use techniques that would benefit from additional RAM during
the competition? No

6) Would you agent benefit from additional disk space? No

One/Two Paragraph Summary of Technique

3 player Kuhn poker bot that calculates expected value of reaching each terminal node from the current decision node and performs the decision that maximises this value. After each hand, the bot updates the probabilities of its opponents decision node with the shown cards. When a card is folded, it is assumed to be weak. 

——-

How should I go about setting up my bot? I have written it using the repository on the website and so it is in python. I could just email you the code if that’s easiest?