This project was prepared to solve the third challenge of "K&H Hack3" Hackathon " https://hack3.crafthub.events/Metaplayers%20challenge_1_K&H%20HACK3%20hackathon.pdf "
The aim is to develop a Leaderboard Elo system and a teammate/enemy finder based on in-game roles and performance.
This challenge has two little bit different parts but both of them are based on statistical database analysis. The hackers should create an "ELO System" ranking each user depending on their performance in each game that is available on esportmanager.com and has an API connected to esportmanager.com. Right now this is PlayerUnknown's Battlegrounds, League of Legends, Teamfight Tactics and Counter-Strike: Global Offensive. But the system's logic should be implemented to other games also, that will be available on the platform.
Leaderboard Tournament based on statistics Important to note that the tournament itself should not be developed. The main task is to develop and create an ELO system and calculating method for the leaderboard.
With the use of the "ELO System'' we would like to create monthly/weekly/annual leaderboard tournaments. The user joins the monthly/weekly/annual tournament and all they have to do is play games as usual. The platform should track the users' all played games and statistics. Within this 30 days the user plays as much as they want but only the 10 best games count in the tournament. Based on these 10 games the total score is calculated, the first 16, 32, etc users are rewarded. Regarding the scoring it depends on the game but usually the following statistics count: win, kill count, death count (it counts as minus point), assists, farm, healing, damage dealt, etc. These attributes should be changed for each game.
Creating the leaderboard for one game is a must (preferably it should be League of Legends) but having more games is a plus (for example: CS:GO, PUBG, TFT).
Teammate/enemy finder based on in-game statistics a roles - queue and recommendation Using the same "ELO System" and statistic database of the players' performance the platform recommends teammates/duo partners play with or enemies to play against. For example it considers the in-game rank, in-game role, character usage, etc and finds you the perfect partner for a duo. These attributes should be changed for each game.
There should be a pool of players who want to participate in the "Finder Queue for Teammates" or "Finder Queue for Enemies".