@ihs7/ts-elo
v1.0.0
Published
Elo ranking system implemented in TypeScript
Downloads
39
Readme
@ihs7/ts-elo
ELO calculation library in TypeScript
Supports heads up, team and multiplayer calculations
Installing ts-elo
npm i @ihs7/ts-elo
Two player scenario
Create and calculate match between two players
- Player 1 has 1200 rating and wins
- Player 2 has 1320 rating and loses
- Get results
const match = new Duel();
match.addPlayer(new Player("player-identifier-1", 1200), true);
match.addPlayer(new Player("player-identifier-2", 1320), false);
const results = await match.calculate();
Three player scenario
Create a match between three players
- Player 1 has 1280 rating and places 1st
- Player 2 has 1300 rating and places 2nd
- Player 3 has 1220 rating and places 3rd
- Get results
const match = new FreeForAll();
match.addPlayer(new Player("player-identifier-1", 1280), 3);
match.addPlayer(new Player("player-identifier-2", 1300), 2);
match.addPlayer(new Player("player-identifier-3", 1220), 1);
const results = await match.calculate();
The calculations is based on:
- Player 1 won Player 2 and Player 3
- Player 2 won Player 3 and lost to Player 1
- Player 3 lost to Player 1 and Player 2
Two versus two scenario
- Team 1 consists of Player 1 with 1230 ELO and Player 2 with 1260 ELO
- Team 2 consists of Player 3 with 1120 ELO and Player 4 with 1410 ELO
- Team 1 wins Team 2
- Get results
const match = new TeamMatch();
const team1 = match.addTeam("1", 2);
team1.addPlayer(new Player("player-identifier-1", 1230));
team1.addPlayer(new Player("player-identifier-2", 1260));
const team2 = match.addTeam("2", 1);
team2.addPlayer(new Player("player-identifier-3", 1120));
team2.addPlayer(new Player("player-identifier-4", 1410));
const results = await match.calculate();
Each team has a rating which is an average of the team members.
- Team 1 has a rating of 1245 ((1230+1260)/2)
- Team 2 has a rating of 1265 ((1120+1410)/2)
Calculation Strategies in Team Matches
By default when creating TeamMatch, the results are calculated using average team rating as if the teams were individuals. However, you can also use WEIGHTED_TEAMS
strategy to have the results calculated based on individual player ratings within the team.
- Team 1 consists of Player 1 with 1000 ELO and Player 2 with 1400 ELO
- Team 2 consists of Player 3 with 1200 ELO and Player 4 with 1600 ELO
- Team 1 wins Team 2
- Get results
const match = new TeamMatch();
const team1 = match.addTeam("1", 2);
team1.addPlayer(new Player("player-identifier-1", 1000));
team1.addPlayer(new Player("player-identifier-2", 1400));
const team2 = match.addTeam("2", 1);
team2.addPlayer(new Player("player-identifier-3", 1200));
team2.addPlayer(new Player("player-identifier-4", 1600));
const results = await match.calculate({
calculationStrategy: CalculationStrategy.WEIGHTED_TEAMS,
});
Each team has a rating which is an average of the team members. This is used as a basis of calculation:
- Team 1 has a rating of 1200 ((1000+1400)/2)
- Team 2 has a rating of 1400 ((1200+1600)/2)
Then after calculating the expected score of each player in the team, the results are calculated based on individual player ratings within the team. For example, Player 1 has a weight of 1000/(1000+1400) ~ 0.4167 and Player 2 has a weight of 0.5833. The same is done for Team 2 and the ELO changes are distributed based on these weights.
With this calculation method, the players are rewarded or penalized based on their rating contribution to the team.
Get expected score between two ratings
Say you have two players, one with 1460 ELO and another with 1130 ELO, and want to know the likelihood of one winning another.
const player1 = new Player("player-identifier-1", 1460);
const player2 = new Player("player-identifier-2", 1130);
const expectedScore = player1.expectedScoreAgainst(player2);
This returns a float number and in this specific scenario the value is 0.8698499 meaning the higher ranked player is estimated to win 87% of encounters.