August 15, 2015

GKRandomDistribution, GKShuffledDistribution and the GKGaussianDistribution

GameplayKit brings random numbers generation among other useful things. So what’s the difference between GKRandomDistribution, GKGaussianDistribution and GKShuffledDistribution? As the name suggests, GKShuffledDistribution is useful for when you want to simulate card deck behaviour. Pulling a random card one after another. Until you go through entire deck, there is no way you will see the same card twice (putting magic tricks aside). GKGaussianDistribution is targeting problem of generating random numbers around the center of the range (2 in the example below). Here is a little sample, illustrating the differences:

import Cocoa
import GameplayKit

let randoms = NSCountedSet()
let shuffles = NSCountedSet()
let gausians = NSCountedSet()

let randomD3 = GKRandomDistribution(lowestValue: 1, highestValue: 3)
let shuffledD3 = GKShuffledDistribution(lowestValue: 1, highestValue: 3)
let gausianD3 = GKGaussianDistribution(lowestValue: 1, highestValue: 3)
for i in 0..<30 {

    print("\($0) \(randoms.countForObject($0))")

    print("\($0) \(shuffles.countForObject($0))")

    print("\($0) \(gausians.countForObject($0))")

Here is the output it produces:

1 14
2 11
3 5

1 10
2 10
3 10

1 2
3 2
2 26

And a visual version:

Although GKRandomDistribution instance produces equally distributed random numbers, it can result in so called “lucky streaks” - getting same value several times in a row. That’s exactly when you want to use GKShuffledDistribution. As for GKGaussianDistribution, it targets cases when you need to generate random numbers around center, not getting too close to the upper and lower bounds (1 and 3 in our example).

Very welcomed additions to Cocoa family.

Where to go from here: