Elo_rating: Elo rating
In AkselA/R-ymse: Ymse (Various)

Calculate updated Elo ratings based on existing rating and results from matches

1
2
3

elo_upd(ra, rb, score, k = 16, sum = length(rb) > 1)

elo_upd_pw(ra, rb, score, k = 16)

`ra`	rating of player A
`rb`	rating of player B
`score`	score respective to player A. Numeric or character; 1, 0.5, 0; win, tie, loss
`k`	a measure of how big the correction should be
`sum`	calculate the new rating based on sum of scores and opponents ratings

# as in example from
# https://en.wikipedia.org/wiki/Elo_rating_system#Mathematical_details
# per 2019-10-04
ra <- 1613
rb <- c(1609, 1477, 1388, 1586, 1720)
score <- c(0, 0.5, 1, 1, 0)
elo_upd(ra, rb, score, k=32, sum=FALSE)
elo_upd(ra, rb, score, k=32, sum=TRUE)

elo_upd_pw(c(1400, 1500), c(1300, 1400), c("w", "t"))

results <- read.table(text="
Player1   Player2   Result
Alice     Bob       Win
Charlie   Dennis    Loss
Elena     Frank     Loss
June      Rashida   Tie", header=TRUE, stringsAsFactors=FALSE)

scores <- read.table(text="
Player    Score
Alice     1150
Charlie   1150
Frank     1150
Bob       800
Dennis    800
Elena     800
June      900
Rashida   1100", header=TRUE, stringsAsFactors=FALSE)

rownames(scores) <- scores$Player

r2 <- results
r2[,1:2] <- scores[as.matrix(r2[,1:2]), 2]

r2u <- elo_upd_pw(r2)

scores.new <- data.frame(Score=c(r2u))
rownames(scores.new) <- as.matrix(results[,1:2])

scores.new <- round(scores.new[rownames(scores),, drop=FALSE])
scores.new$diff <- scores.new$Score - scores$Score
scores.new