Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

folks,

i am new to this forum and not a math expert. so please bear with me if am asking silly questions.

The question is "probability of getting 50 heads from tossing a coin 100 times".

So the answer for this is, I guess, ${100 \choose 50} (2 ^{-100})$.

So all am trying to get is easier way to calculate ${100 \choose 50}$, or another approach to the parent problem only.

Thanks all, appreciate that.

raj

share|improve this question
    
I guess you mean 2^{-100}, but you are correct. To estimate this number you can use Stirling's formula: en.wikipedia.org/wiki/Stirling's_approximation –  Qiaochu Yuan Nov 29 '10 at 14:15
    
oh yes.. thanks for poiting out the typo... –  Rajan Nov 30 '10 at 7:48
add comment

3 Answers

up vote 5 down vote accepted

The coefficient ${2n \choose n}$ for $n$ large can be approximated well (using Stirling's formula) by ${2n \choose n} \approx 4^n / \sqrt{n \pi} $. In particular, letting $n=50$ gives ${100 \choose 50} \approx 4^{50}/ \sqrt {50 \pi}$.

share|improve this answer
    
Another approach is approximating the binomial with a normal distribution, i.e. N(np,np(1-p))=N(50,25) and interpolating the desired value from a standard normal distribution. –  karakfa Jun 13 '12 at 20:58
add comment

See How to calculate binomial probabilities for how to calculate these probabilities while avoiding numerical overflow or underflow.

Also, I believe you meant to type 2^-100 instead of 2^-10 in your question.

share|improve this answer
    
yes John, thanks for pointing out the typo... –  Rajan Nov 30 '10 at 7:48
add comment

The numbers in question here, of course, can be computed exactly. For example, using bignum or GAP (or even WolframAlpha -- the exactly link won't work on here, but I'm sure you can type in "(100 choose 50)*2^(-100)" yourself).

On my home computer, in GAP, it took less than a millisecond. To make it more interesting, I also computed ${100000 \choose 50000} \cdot 2^{-100000}$, which took a bit more than 17 seconds.

gap> Binomial(100,50)/2^100;
12611418068195524166851562157/158456325028528675187087900672
gap> time;
0
gap> Binomial(100000,50000)/2^100000;
<<an integer too large to be printed>>/<<an integer too large to be printed>>
gap> time;
17266

In fact, provided the coin has probability 1/2, the probability will always have a terminating decimal expansion (since binomial coefficients are integers, and 2 divides 10). Here it is in this case:

0.0795892373871787614981270502421704614029315404247333213573478705171737601631321012973785400390625
share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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