4/3/2023 0 Comments True game time decision![]() With this item they are then strong enough to obtain more kills and so on until they can lead their team to a win. ![]() For example, a player obtaining the first kill of the game nets them gold that can be used to purchase more powerful items. Modelling the events in this way is even more important in games such as League of Legends as taking objectives and kills lead towards both an item and level advantage. The aim of research such as this one mentioned is to provide more detailed insight beyond a simple box score (number of points or kill gained by player in basketball or video games respectively) and consider how teams perform when modelled as a sequence of events connected in time. Sequences like this being modelled statistically is nothing new for years now researchers have considered how this is applied in sports, such as basketball ( ), where a sequence of passing, dribbling and foul plays lead to a team obtaining or losing points. ![]() We therefore have a sequence of events dependent on previous events that lead to one team destroying the other’s base and winning the game. Gaining an advantage enables the players to become stronger (obtain better items and level up faster) than their opponents and, as their advantage increases, the likelihood of winning the game also increases. League of Legends is a team oriented video game where on two team teams (with 5 players in each) compete for objectives and kills. This is very much a work in progress and is only designed to introduce the idea of what can be achieved if more complex machine learning methods are introduced into the game that go beyond simple summary statistics as shown in the image below. I have included the first two parts to give some clarity in the reasoning behind my final decisions on how the environment is modelled. I have made each part available on Kaggle to give a better understanding of how the data is processed and the model is coded. Source This three part project aims to model League of Legends matches into Markov Decision Processes then apply Reinforcement Learning to find the optimal decision that also account for a player’s preference and go beyond simple ‘score board’ statistics.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |