Elo ranking system for Go (Golang) programming language.
Elo ranking system is a system for ranking players in a competitive environment, by calculating expected outcome in a match between two ranked players and their new ranks based on the outcome of a match. Can be used in head-to-head mathes, as well as team based matches.
There are two factors that affect how Elo works - K-factor and deviation.
Deviation controls how much the difference in ratings affects the expected outcome. The higher the number, the higher expected chance that a better rated player wins. Default deviation is 400.
K-factor controls how much points are gained/lost after the outcome of the match is known. The higher the factor, players rating is more affected. This is especially useful if you want to control point fluctuation based on number of games played. Players with less games played should have higher k-factor, because their early ranking doesn't necessarily replicate their skill level. Default k-factor is 24.
rankA := 1500
rankB := 1600
elo := NewElo() // or NewEloWithFactors(k, d) for custom factors
// Expected chance that A defeats B
// use ExpectedScoreWithFactors(rankA, rankB, deviation) to use custom factor for this method
elo.ExpectedScore(rankA, rankB) // 0.3599350001971149
// Results for A in the outcome of A defeats B
score := 1 // Use 1 in case A wins, 0 in case B wins, 0.5 in case of a draw
elo.RatingDelta(rankA, rankB, score) // 20
elo.Rating(rankA, rankB, score) // 1520
outcomeA, outcomeB := elo.Outcome(rankA, rankB, score)
outcomeA.Delta // 20
outcomeA.Rating // 1520
outcomeB.Delta // -20
outcomeB.Rating // 1580
All Elo methods have WithFactors variants, where you can override factors for that specific method call:
kFactor = 40
deviation := 800
elo := NewEloWithFactors(kFactor, deviation) // set default factors on elo struct
// override factors for indiviual methods
elo.ExpectedScoreWithFactors(rankA, rankB, deviation)
elo.RatingDeltaWithFactors(rankA, rankB, score, kFactor, deviation)
elo.RatingWithFactors(rankA, rankB, score, kFactor, deviation)
elo.OutcomeWithFactors(rankA, rankB, score, kFactor, deviation)