Sorting under the influence of `undefined`

results.txt

Original
[ { id: 1, created: 300 },
  { id: 2, created: 200 },
  { id: 3, created: 500 },
  { id: 4 },
  { id: 5, created: 400 },
  { id: 6, created: 100 },
  { id: 7 } ]

Naive
[ { id: 2, created: 200 },
  { id: 1, created: 300 },
  { id: 3, created: 500 },
  { id: 4 },
  { id: 6, created: 100 },
  { id: 5, created: 400 },
  { id: 7 } ]

Robust
[ { id: 2, created: 200 },
  { id: 1, created: 300 },
  { id: 3, created: 500 },
  { id: 4 },
  { id: 6, created: 100 },
  { id: 5, created: 400 },
  { id: 7 } ]

sorting.coffee

# Sort functions
naive = (a, b) ->
    a.created - b.created

robust = (a, b) ->
    r = a.created - b.created
    if isNaN r
        r = 0
    r

# List
list = [
    {id: 1, created: 300},
    {id: 2, created: 200},
    {id: 3, created: 500},
    {id: 4},                # missing key
    {id: 5, created: 400},
    {id: 6, created: 100},
    {id: 7}                 # another missing key
]


# Original
console.log '\nOriginal'
console.log list

# Naive
console.log '\nNaive'
l2 = list.slice(0)
l2.sort naive
console.log l2

# Robust
console.log '\nRobust'
l3 = list.slice(0)
l3.sort robust
console.log l3

The robust output seems to be the same as the naive one. Is that indeed the case or was this just a typo? =)