3
$\begingroup$

I recently took a HackerRank test for a Data Science position and got the following question wrong:

A casino has 10 slot machines. Eight of them are "Wheel of Fortune" themed, and two are "Jeopardy" themed. However, you do not know which machines are which. A "Wheel of Fortune" slot machine wins on 20% of spins which a "Jeopardy" machine wins on 50%. Say you play a round and are a winner. What is the probability that this is a "Wheel of Fortune" machine?

I used posterior probability to calculate this: WoF = Wheel of Fortune

J = Jeopardy

W = Win

L = Lost

P(WoF|W) = (P(W|WoF) * P(WoF)) / P(W)

P(WoF) = 80/100 = 4/5
P(W|WoF) = 20/100 = 1/5
P(W) = (20/100)*(8/10)+(50/100)*(2/10) = 13/50

P(WoF|W) = 8/13

I came up with the answer of 8/13. Can anyone help me identify what I did wrong?

$\endgroup$
  • 1
    $\begingroup$ Why do you believe that you did something wrong? Do the HackerRant Test answer sheets give a different result? $\endgroup$ – Dilip Sarwate Sep 22 '16 at 3:22
2
$\begingroup$

I am not convinced you did anything wrong here. Having looked at the work provided I suspected the answer was correct, so I went ahead and made a quick simulation in Python to test the answer you provided (8/13 or ~0.6154). Note that this simulation assumes the probability of picking a machine to play is uniform across all machines.

import random

machine_list = [1,1,0,0,0,0,0,0,0,0]

def pick():
    p = random.randint(0,9)
    return machine_list[p]

def play():
    p = pick()
    if p == 1: #J
       w = random.randint(1, 2)
        if w == 1:  # win
            return 3
        if w == 2:  # lose
            return 4
    if p == 0: #WoF
        w = random.randint(1, 10)
        if w <= 2:  # win
            return 1
        else:  # lose
            return 2


winlist = []
woflist = []

for i in range(10000000):
    p = play()
    if p == 1 or p == 3:
        winlist.append(p)
    if p == 1:
        woflist.append(p)

print len(woflist) / float(len(winlist))

With the following output:

0.615330464637
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.