|Institution:||KTH Royal Institute of Technology|
|Keywords:||Natural Sciences; Computer and Information Science; Computer Science; Naturvetenskap; Data- och informationsvetenskap; Datavetenskap (datalogi)|
|Full text PDF:||http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-166310|
The game of Yahtzee is a semi-strategic luck based game, which means it should be possible to maximize the score with an optimal strategy. The main purpose of the study is to compare the results of the optimal algorithm with other useable strategies when playing Yahtzee. To receive interesting results, the strategies and decisions made by the algorithms will be compared and analyzed in performance as well as by creating an environment where they have to choose from the same set of dices. To further see how well the different algorithms performed, Human Trials were conducted where 6 humans contributed by playing the game of Yahtzee 300 times. These test subjects were familiar with the game of Yahtzee and in this study it is concluded that these subjects had through reinforcement learning created an almost optimal play style. Our conclusion is that the Optimal Algorithm performs better than other algorithms that because it does not take any risks while playing while it tries to maximize the score but doing this uses a great amount of computation power, approximately 13.8GB of RAM and around 22 hours to complete.