The Great Cherry Ripe Scandal, Part II

First we set up the chocolate randomisation process, but this time making 25% of the bars Cherry Ripe:

chocolate_types<-c("Cherry Ripe","Dairy Milk","Crunchie", "Caramello","Flake")

make_pack<-function(){
  sample(chocolate_types, 8, replace=TRUE, prob=c(25,19,19,18,19))
}

make_pack()

## [1] "Flake"       "Flake"       "Caramello"   "Dairy Milk"  "Dairy Milk" 
## [6] "Caramello"   "Cherry Ripe" "Crunchie"

make_pack()

## [1] "Flake"       "Cherry Ripe" "Caramello"   "Dairy Milk"  "Cherry Ripe"
## [6] "Cherry Ripe" "Cherry Ripe" "Caramello"

replicate(5, make_pack())

##      [,1]          [,2]          [,3]          [,4]          [,5]         
## [1,] "Cherry Ripe" "Dairy Milk"  "Cherry Ripe" "Caramello"   "Cherry Ripe"
## [2,] "Cherry Ripe" "Crunchie"    "Dairy Milk"  "Crunchie"    "Crunchie"   
## [3,] "Dairy Milk"  "Cherry Ripe" "Crunchie"    "Flake"       "Cherry Ripe"
## [4,] "Flake"       "Dairy Milk"  "Flake"       "Cherry Ripe" "Caramello"  
## [5,] "Cherry Ripe" "Cherry Ripe" "Cherry Ripe" "Cherry Ripe" "Dairy Milk" 
## [6,] "Flake"       "Caramello"   "Flake"       "Cherry Ripe" "Cherry Ripe"
## [7,] "Cherry Ripe" "Dairy Milk"  "Dairy Milk"  "Cherry Ripe" "Crunchie"   
## [8,] "Cherry Ripe" "Cherry Ripe" "Cherry Ripe" "Flake"       "Flake"

Now, we need to work out which bar we have the most of, and count them

count_most<-function(a_pack){
    count <- table(a_pack)
    most <- which.max(count)
    count[most] 
}

a_test_pack<-make_pack()
a_test_pack

## [1] "Crunchie"    "Caramello"   "Flake"       "Cherry Ripe" "Caramello"  
## [6] "Flake"       "Flake"       "Crunchie"

count_most(a_test_pack)

## Flake 
##     3

We’re going to do this lots of times, so we need a function that asks how many times and then does it:

make_and_count_most<-function(how_many){
    replicate(how_many,{
        pack<-make_pack()
        count_most(pack)
    })
}

make_and_count_most(3)

## Crunchie    Flake Crunchie 
##        3        3        4

And, finally, we take advantage of the power of modern computing by running this a whole heap of times. For curiosity – because the code was written to be clear, not to be fast – I’ll also see how long it takes on my three-year old laptop.

a_whole_heap <- 1e5
system.time(results <- make_and_count_most(a_whole_heap))

##    user  system elapsed 
##  13.319   0.032  13.352

table(results)

## results
##     2     3     4     5     6     7     8 
## 15230 55283 23510  5159   739    76     3

table(results[names(results)=="Cherry Ripe"])

## 
##     2     3     4     5     6     7     8 
##  4784 17942  8713  2336   420    47     3