Python 3.3: Birthday Probability

You should create the list like this. test isn't a very useful name. Consider something like make_birthday_list instead

def test(count):
    return [random.randint(1, 365) for x in range(count)]

The test in duplicates already evaluates to a bool, so you can just do this. It's not a good idea to use l as a variable name (looks too much like 1) so I changed it to the_list

def duplicates(the_list):
    return len(the_list)!=len(set(the_list))

You'll have to run the test over and over to get the probability

eg.

num_samples = 10000
for count in range(100):
    dup = 0
    for test_number in range(num_samples):
        the_list = make_birthday_list(count)
        if duplicates(the_list):
            dup += 1
    print(count, dup/num_samples)   # / returns a float in Python3

test does not return a list of length count. The result only has one element. You would want to make use of list comprehension to get this to work nice:

[random.randint(1, 365) for x in range(count)]

Then for step 3:

def probability(count,num):
    dup_trues = sum([duplicates(test(count)) for i in range(num)])  #***
    dup_falses = num-dup_trues
    return float(fails)/num

The host important line in the above function is labeled #*** Here's a breakdown of how it works:

The stuff in the square brackets is a list comprehension.

for i in range(num) says do this stuff with i=0, i=1, ...i=num-1. You know what duplicates and test does. So the list ends up with a bunch of trues and falses (ones and zeros)

sum adds up the elements of the list. So this gives us the number of trues

The next line dup_falses = num-dup_trues says we have num results altogether, those that aren't true are false.

return float(fails)/num the float is not necessary in python 3