6
$\begingroup$

Suppose you have a bag of 100 coins of which 1 is biased with both sides as Heads. You pick a coin from the bag and toss it three times. The result of all three tosses is Heads. What is the probability that the selected coin is biased?

My answer:-

P(selecting a biased coin) = 1/100
P(getting a head thrice with the biased coin) = 1

P(selecting an unbiased coin) = 99/100
P(getting a head thrice with the unbiased coin) = 1/8

P(selecting a biased coin|coin toss resulted in 3 heads) = 
          P(selecting a biased coin and getting heads thrice)/
          [P(selecting a biased coin and getting heads thrice) +
           P(selecting an unbiased coin and getting heads thrice)] 

= (1/100)/[(1/100) + (99/800)]
= (1/100)/(107/800)
= 8/107
= 0.0747

Is this correct? Thanks.

$\endgroup$
1
  • $\begingroup$ @CarlosAccioly This question was posed to me in an examination on Sunday. Just wanted to verify my answer. $\endgroup$
    – tejas_kale
    Feb 19, 2013 at 12:44

3 Answers 3

8
$\begingroup$

Your answer is right. The solution can be derived using Bayes' Theorem:

$P(A|B) = \frac{P(B|A)P(A)}{P(B)}$

You want to know the probability of $P(\text{biased coin}|\text{three heads})$.

What do we know?

There are $100$ coins. $99$ are fair, $1$ is biased with both sides as heads.

With a fair coin, the probability of three heads is $0.5^3 = 1/8$.

The probability of picking the biased coin: $P(\text{biased coin}) = 1/100$.

The probability of all three tosses is heads: $P(\text{three heads}) = \frac{1 \times 1+ 99 \times \frac{1}{8}}{100}$.

The probability of three heads given the biased coin is trivial: $P(\text{three heads}|\text{biased coin}) = 1$.

If we use Bayes' Theorem from above, we can calculate

$$P(\text{biased coin}|\text{three heads}) = \frac{1 \times 1/100}{\frac{1 + 99 \times \frac{1}{8}}{100}} = \frac{1}{1 + 99 \times \frac{1}{8}} = \frac{8}{107} \approx 0.07476636$$

$\endgroup$
1
  • 2
    $\begingroup$ Cool! Cool! Cool! $\endgroup$
    – Behacad
    Feb 19, 2013 at 14:33
1
$\begingroup$

Here is a write that describes something very similar to that. The Bayes approach is the right way to proceed.

$\endgroup$
0
$\begingroup$

In general, P(biased coin|k heads) = (2^k)/[(2^k) + 99] Where k is no of consective heads

so if the trick coin was tossed 3 times (2^3)/[(2^3) + 99] = 8/(8+99) = 8/107 = 0.07

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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