Manually setting the Tail-Sensitive Confidence Bands from Aldor-Noiman (2013) in R

Question

I'm trying to implement the Aldor-Noiman's algorithm in order to visualize the Tail-Sensitive (TS) confidence bands using R. But I'm missing something at the time of calculate the tail probability A_{i} and find the minimum C^{m}.

The main idea of the algorithm is this

This is my code

ts_confidence_intervals <- function(M, n, alpha){
  mu <- 0
  sigma <- 1
  CValues <- numeric(M)
  LowerBound <- numeric(n)
  UpperBound <- numeric(n)
    
  for(i in 1:M){
    samples <- rnorm(n, mu, sigma)
    
    Zi <- (samples - mu)/sigma
    
    Yi <- qnorm(pnorm(Zi))
    
    Y <- sort(Yi)
    
    for(i in 1:n){
      A <- qbeta(Y[i], i, n + 1 - i)
    }
    
    C <- 2*min(min(A, 1 - A))
    
    CValues[i] <- C 
    
  }
  
  gamma <- quantile(CValues, 1 - alpha)
  
  for(i in 1:n){
    LowerBound[i] <- qnrom(qbeta(gamma/2, i, n + 1 - i)) 
    UpperBound[i] <- qnrom(qbeta(1 - gamma/2, i, n + 1 - i)) 
  }
  
  return(list("LowerBound" = LowerBound, "UpperBound" = UpperBound))
  
}

result <- ts_confidence_intervals(500, 10, 0.05)
print(result$LowerBound)
print(result$UpperBound)