# Read data:
a<- read.table("http://www.stat.ucla.edu/~nchristo/datac183c283/statc183c283_5stocks.txt", header=T)

#Convert prices into returns:
r1 <- (a$P1[-length(a$P1)]-a$P1[-1])/a$P1[-1] 
r2 <- (a$P2[-length(a$P2)]-a$P2[-1])/a$P2[-1] 
r3 <- (a$P3[-length(a$P3)]-a$P3[-1])/a$P3[-1] 
r4 <- (a$P4[-length(a$P4)]-a$P4[-1])/a$P4[-1] 
r5 <- (a$P5[-length(a$P5)]-a$P5[-1])/a$P5[-1]

#Use stocks 1, 4, 5:
ret <- data.frame(r1,r4,r5)

row.names(ret) <- a$date[-216]

#Convert the data frame into a matrix:
ret <- as.matrix(ret)

#Load the package:
library(stockPortfolio)

#Choose a model:
port_model <- stockModel(ret, model="none", Rf=0.001)

#Optimize:
op <- optimalPort(port_model)

#Portfolio possibilities curve, cloud of points, etc.
portPossCurve(port_model, xlim=c(0,0.20), ylim=c(-0.0008,0.007))

portCloud(port_model, add=TRUE)

points(op, pch=19, add=TRUE)
points(op$risk, op$R, pch=19, col="green")

#Draw the tangent:
Rf <- 0.001
segments(0, Rf, op$risk, op$R)

#If you want to extend the tangent beyond G:
slope <- (op$R-Rf)/op$risk
segments(0, Rf, 1.4*op$risk, Rf+slope*1.4*op$risk)

#Place the portfolio 60% G + 40% Rf on the CAL:
points(0.60*op$risk,0.6*op$R+0.4*0.001)