Index selection on 2 variables, 1 good & bad

library(mvtnorm)
library(MBESS)
rgMatrix <- matrix(ncol=2, c(1,-0.3, -0.3, 1))
covMatrix <- cor2cov(rgMatrix, sd=rep(1,2))
generate <- function(mu=c(0,0), n=1000) { rmvnorm(n, mean=mu, sigma=covMatrix, method="svd") }

df <- as.data.frame(generate())
colnames(df) <- c("Good", "Bad")

df$Index <- (df$Good * 1) - (df$Bad * 0.5)
cutoff <- quantile(df$Index, probs=0.90)
df$Selected <- df$Index > cutoff
library(ggplot2)
qplot(Good, Bad, color=Selected, data=df) + geom_point(size=5)