Here are three data sets collected over the past two years: 2017 COMP717 Arti?cial Intelligence https://goo.gl/GTbDsL 2018 COMP717 https://bit.ly/2HQDDMi 2018 COMP808 Logic, Games and Automata http://bit.ly/2vahHmZ For COMP717, there are in total 3 labs (A, B, and C, in chronicle order). The participants in B and C were informed the winners choices of earlier sessions, apart from that all the rules were the same. For COMP808, there is only 1 lab and the game is repeated with slightly changed rules (note that there was one student less (S12) in the second game). To respect privacy, we replaced the name and student IDs with pseudo IDs of a letter followed by a number. Note that under the [File] menu, it can be downloaded for o?ine access:

Your task is to study these data ?les and write a report of 3 pages maximum (longer reports will be penalized by a direct subtraction of 1% for every extra line over 10 pages). You need to discuss at least the following points: (A) what is rationality? Were the winners win by rational thinking, pure luck, or a combination of both? Your argument should be supported by evidences from the data sets. (b) Give a classi?cation of players in terms of di?erent levels of rationality and their knowledge about others rationality. You need to have clear rules to decide which class a player belongs to. You shall utilize the reasons given by the players. (22 marks) 6 (B) For COMP717s 2017 and 2018 data sets, explain the possible reasons of Lab Bs winning number is lower than that of A, and Lab Cs winning number is lower than that of B. Compare the di?erence between them. (40 marks)

(C) For COMP808 2018 data set, you are able to see the change of number choices for each ID.

Explain the possible reasons of the second games winning number is lower than that of the ?rst game. What impact are shown by the change of rules in the second game? Design an experiment to test or verify your reasons. Recruit some friends or family to play in your experiment (with no less than 6 players). Hint: youll need to think about what reward to be o?ered. We used marks for the papers, but it would not likely give enough incentive to your friends who are not in this paper.

(D) What is the pure strategy Nash equilibrium of this game? None of the lab winners chooses this strategy. If we repeatedly play this game, what would be likely to happen? Test your predications by design an experiment (could integrate with the experiment for (c)). (15 marks) Extra marks (up to 10) will be given if you are able to do something extra, e.g., visualising the data in an interesting and meaningful way, extensive experiments.