rrace() and rrank_single() are the same

[RobinHankin]

Well almost the same. rrace() is higher-level, dealing with named vectors, while rrank_single() is lower-level, dealing with numeric vectors.

Function rrace() is documented at man/rrace.Rd; and function rrank_single() is documented (alongside other functions such as rorder_single() at man/rrank.Rd.

The best solution, I think, is to delete rrace.Rd and add its content to rrank.Rd. And of course rationalise the code, which currently reads

`rrank_single` <- function(p){
    ell <- length(p)
    r <- integer(ell)
    i <- 1   # 'i' means placing.
    while(any(p > 0)){  
        m <- sample(ell,1,prob=p) #randomly chosen competitor
        r[i] <- m   # place 'i' is competitor 'm'  
        p[m] <- 0   # competitor 'm' can't place again.
        i <- i+1    # consider the next place
    }
    return(r)
}

[snip]

`rrace` <- function(strengths){
    out <- strengths
    for (i in seq_along(out)) {
        nextfinisher <- sample(names(strengths),size=1,prob=strengths)
        out[i] <- nextfinisher
        strengths <- strengths[names(strengths) != nextfinisher]
    }
    names(out) <- NULL
    return(out)
}

Looking at the above, it is not clear which is better, or faster. they use very different methods. But either way, this is not following the DRY rule unless