Shuffle: Fisher-Yates vs Naive Shuffles (Swift 4)

SimpleShuffle.swift

extension Array {
    
    /// Simple implementation of Fisher-Yates
    
    mutating func shuffle() {
        for i in 0 ..< count - 1 {
            let j = i + Int(arc4random_uniform(UInt32(count - i)))
            swapAt(i, j)
        }
    }
    
    /// Naive implementation that introduces biases
    
    mutating func shuffleBiased() {
        for i in 0 ..< count {
            let j = Int(arc4random_uniform(UInt32(count)))
            swapAt(i, j)
        }
    }
}

The naive shuffle algorithm generates a very biased result. For example, shuffling an array of 10 numbers 10,000,000 times resulted in a distribution of:

Whereas Fisher-Yates algorithm results in the following distribution:

Pursuant to https://stackoverflow.com/questions/46615663/shuffle-an-array