Birth Certificate!

In 2013, Hospital A delivered 200 \large 200 babies and Hospital B delivered 400 \large 400 babies. Given that the probability of having a boy and a girl is equal, which hospital has a greater chance of delivering equal number of boys and girls?

Both have 100% chance Hospital B Hospital A Both have equal chances.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

15 solutions

Satvik Golechha
Apr 13, 2014

If there were only 2 deliveries, the probability of equal number of boys and girls would be 0.5.

If there were 200 deliveries, then for getting a half probability, we will need 100 and 99 boys and girls before the last one, and its probability will definitely be less than half.

So we get the general case- More the number of deliveries, lesser the chance of equal number of bays and girls. And so is the answer- Hospital A.

It does sound counter-intuitive doesn't it? We generally expect probabilities to "even out" in the long run.

A potential good followup question could be "Which hospital has a greater chance that the number of boys delivered differs from the number of girls delivered by less than 20?"

Calvin Lin Staff - 7 years, 1 month ago

Log in to reply

Exactly, the larger the data set, the more accurately it should show the odds. AKA evening out. Hence why I chose Hospital B.

Dorian Thiessen - 7 years, 1 month ago

Log in to reply

There is a big difference between "the odds evening out", and the "probability of exactly that event".

When we say that "the odds even out", what we mean is that it is more likely that 49-51% (or any range around 50%) of the babies born are boys. However, because there are much more specific events, the probability of each event is actually lower. For example, in hospital A, we included having 98, 99, 100, 101, 102 boys (5 events), while in hospital B we included having 196, 197, 198 199, 200, 201, 202, 203, 204 boys (9 events)

We know that the probability of having 2 n 2n children of which n n are boys an n n are girls is ( 2 n n ) × 0. 5 n × 0. 5 n { 2n \choose n } \times 0.5^ n \times 0.5^n . With stirling approximation of n ! 2 π n ( n e ) n n! \sim \sqrt{2\pi n } \left( \frac{n}{e} \right)^n , we can approximate this probability to be

2 π 2 n ( 2 n e ) 2 n ( 2 π n ( n e ) n ) 2 × 0. 5 n × 0. 5 n = 1 π n \frac{ \sqrt{2\pi 2n } \left( \frac{2n}{e} \right)^{2n} } { \left( \sqrt{2\pi n } \left( \frac{n}{e} \right)^n \right)^2} \times 0.5^{n} \times 0.5^{n} = \frac{1 } { \sqrt{\pi n} }

The probability of this specific event decreases as n n increases.

.

Calvin Lin Staff - 7 years, 1 month ago

Log in to reply

@Calvin Lin Sir, this means that as n approaches infinity, the probability tends to zero?

vatsal shah - 5 years, 9 months ago

that is what my views are.

Sadiq Khan - 7 years, 1 month ago

You are correct. I, too, was initially thinking about the Normal Distribution Curve (taking the left side of the bell as extra girls/boys and the right side as the other one) but then, again, I think that comes into effect only if the number is really large eg: 1 million or something. Normal distribution can't be applied in small-ish cases such as 200-400. That is why I went with A - strict simple probability. However I do think that the author takes note of this potential issue in the future and try to make sure that he keeps the numbers low enough not to make a difference (eg: Hospital A has 20 and Hospital B has 35).

Srujan Gupta - 7 years, 1 month ago

Log in to reply

That interpretation is not correct. See the above comment that I just made.

Normal approximations to the binomial distribution happen when n > 50 n>50 . (of course, this range depends on your limits of acceptable error). With p = 50 p=50 , the value of n n can go much lower.

Calvin Lin Staff - 7 years, 1 month ago

Log in to reply

@Calvin Lin Ah. Fair enough. Thanks :D

Srujan Gupta - 7 years, 1 month ago

But more the events, the more would be the Probability of getting equal boys and girls ! Couldn't understand

Soutrik Bandyopadhyay - 7 years, 1 month ago

Log in to reply

The Law of Large: The more people we survey, the more accurate the result. Larger samples are better

Lakshay Singhal - 7 years, 1 month ago

That is precisely the point. We are only measuring 1 event, namely the one where there are exactly n n boys and exactly n n girls. We are not measuring the events where the number of boys and the number of girls are approximately the same. Because we need exactly the same, we only have 1 event, whose probability decreases.

Calvin Lin Staff - 7 years, 1 month ago

Log in to reply

Ohhh snap. I didn't understand your reply to me but that comment cleared it up for me. You are certainly correct. There are more possible combinations in group of 400 children than in the group of 200 children.

Dorian Thiessen - 7 years, 1 month ago

yes thee law of large is suitable when the more people we collect to study more accurate chances to get equality

Mo Hsin - 7 years, 1 month ago

i think this answer is wrong because if you select any hospital its 1/2 and in each hospital the probability to get the required outcome is 1/2 (it may happen or may not happen ) hence when we select a as our result by bayers therom 1/2.1/2/1/2.1/2+1/2.1/2 and for b its the same thing so the result will be they are eaual

karan khosla - 7 years, 1 month ago

Nice solution same approach.

Mardokay Mosazghi - 7 years, 1 month ago

If there were only 2 deliveries, the probability of equal number of boys and girls would be 1/3 and not 0.5 since there are 3 possible combinations b/w boys and girls. 1 ) both boys 2) both girls 3) both boy and girl.

Gaurav Garodia - 7 years, 1 month ago

Log in to reply

wrong. there are 4 combinations: 1. boy and boy 2. girl and girl 3. boy and girl 4. girl and boy

  1. and 4. satisfy the given condition. hence P= 2/4 Therefor, P=0.5

Ankan Biswas - 7 years, 1 month ago

Log in to reply

  1. and 4.*, not 1. and 4.

Ankan Biswas - 7 years, 1 month ago

In this case it would be 0.5, as in GG, BB, BG, GB.

Mahesh Gokhale - 7 years, 1 month ago

Well, its quite simple actually, keeping in mind that the probability of male or female is equal...it does not effect which hospital you choose...but statistically speaking...probability of hospital with double amount of deliveries has a bigger chance of getting a probability of 50percent

Khubaib Khawar - 7 years, 1 month ago

The lesser the cases,the lesser the chances of getting he ratio wrong

Raj Sikdar - 7 years, 1 month ago

if there were 2 deliveries, chance of getting equal no of boy & girl would be .33, as this is a combination prob. NOT PERMUTATION. we dont care which order we pick the events. thus, there will be chances like: (2 boy), (2 girl), (1 boy, 1 girl).

Arpan Laha - 7 years, 1 month ago

Log in to reply

No, you missed (1 girl, 1 boy), we will also see the order.

Satvik Golechha - 7 years, 1 month ago

For me all these arguments are tricky. If probability is no more than probability, having same probability to get same number of boys and girls, the answer for me is that no matter how big is the sample, probability is the same. Another thing is what we get once we "throw the dice".

Manuel Doval - 7 years, 1 month ago
Scott Kominers
Apr 19, 2014

If boys and girls are born with identical (independent) probability 1 2 \frac{1}{2} , then there are exactly ( n n / 2 ) \binom{n}{n/2} possible birth orders of boys and girls in a hospital in which n n total children are born. (To see this, note that it suffices to determine which children are boys -- all the others must be girls.)

Meanwhile, there are 2 n 2^n possible birth orders in total, all of which arise with equal probability.

Thus, the probability that a hospital with n n births has an equal number of boys and girls is p ( n ) : = ( n n / 2 ) 2 n . p(n):=\frac{\binom{n}{n/2}}{2^n}.

We can explicitly compute that p ( 200 ) = 11318564332012910145675522134685520484313073709426667105165 200867255532373784442745261542645325315275374222849104412672 0.0563485 , p(200)=\frac{11318564332012910145675522134685520484313073709426667105165}{200867 255532373784442745261542645325315275374222849104412672}\approx 0.0563485, and p ( 400 ) = 12869062516926804121621985040050248344651365672634706029356132446990416546581494824792074439872061618945506074688346765 322781234760863573706989896500376484291213224103652939103832419567580952752105149328705669160017228929487896496593436672 0.0398693 < p ( 200 ) . p(400)=\frac{1286906251692680412162198504005024834465136567263470602935613244699 0416546581494824792074439872061618945506074688346765}{3227812347608635 7370698989650037648429121322410365293910383241956758095275210514932870 5669160017228929487896496593436672}\approx 0.0398693<p(200).

Alternatively, we can observe that the derivative of p ( n ) p(n) is p ( n ) = 2 n ( n n / 2 ) ( H n / 2 H n + log ( 2 ) ) , p'(n)=-2^{-n} \binom{n}{n/2} \left(H_{n/2}-H_n+\log (2)\right), where H k H_k is the k k -th harmonic number. As H n / 2 H n H_{n/2}-H_n (which is negative) is bounded below by log ( 2 ) -\log(2) , we see that p ( n ) < 0 p'(n)<0 -- so that p ( n ) p(n) is decreasing -- for n 0 n\geq 0 ; whence we must have p ( 200 ) > p ( 400 ) p(200)>p(400) .

I have found the new Sonnhard

Daniel Wang - 7 years, 1 month ago

Log in to reply

Yes, yes you did :)

Mursalin Habib - 7 years, 1 month ago

I think the question needs to be worded differently. I read it very quick and selected the obvious without answering the actual question. Given the staff had to reply several times to clear things up explains a lot. Imo....

pud veinz - 6 years, 11 months ago
Narayan Regmi
Apr 19, 2014

You may tempt to give the equal probability of having equal number of boys and girls for both hospitals viz. A and B, as the probability of having a boy and a girl is equal but it differs as the number of possible outcomes increases.

Let me generalize it this way. For example we have 2 babies born. The probability of delivering 1 boy and 1 girl is: a) First boy and second girl OR b) First girl and second boy

So it is: (1/2 * 1/2) + (1/2 * 1/2) = 1/2

Let the number of babies born is 4. Therefore, the probability of delivering 2 boys and 2 girls is a) First boy and second boy and third girl and fourth girl OR

b) First boy and second girl and third boy and fourth girl OR

c) First boy and second girl and third girl and fourth boy OR

d) First girl and second girl and third boy and fourth boy OR

e) First girl and second boy and third girl and fourth boy OR

f) First girl and second boy and third boy and fourth girl OR

So it is : (1/2 * 1/2 * 1/2 * 1/2) + (1/2 * 1/2 * 1/2 * 1/2) + (1/2 * 1/2 * 1/2 * 1/2) + (1/2 * 1/2 * 1/2 * 1/2) + (1/2 * 1/2 * 1/2 * 1/2) + (1/2 * 1/2 * 1/2 * 1/2) = 6/16 = 3/8

i.e, 4 !/(2! * 2!) is the favorable outcome and 2^4 is possible outcome

= 6/16.

So for 200 it will come as (200!/(100! * 100!) / 2^200 and for 400 it will come as (400!/(200! * 200!) / 2^400)

Therefore if the number of babies born are less the probability of having equal boys and equal girls is greater.

Thanks

we dont care if the 1st baby is boy or girl. we dont care the order of babies. so, if there are 4 babies, chance of having equal no of boy,girl would be 1/5.. 5 cases are: (4 boys), (3 boy,1 girl), (2b,2g), (1b,3g), (4g). what u r saying is right when we take the order into account. bt here, we r just concerned about the ultimate result.

Arpan Laha - 7 years, 1 month ago

Thnx a lot I guess ur thread was the best as it explained the problem in the simplest way.

Faraz Ahmed - 7 years, 1 month ago
Mohamed Arafa
Apr 19, 2014

P(A) = (.5^100)(.5^100) = 0.5^200 P(B) = (.5^200)(.5^200) = 0.5^400 Hence, P(A) > P(B)

I think you've forgotten the combination operation in finding the two probabilities, i.e. P(A) = (200C100)(.5^200) and P(B) = (400C200)(.5^400). It doesn't change the answer, anyway.

Clifford Lesmoras - 7 years, 1 month ago

Log in to reply

I cancelled it because it's equal 1

Mohamed Arafa - 7 years, 1 month ago

Log in to reply

no. 200C100 and 400C200 are not equal.

Ron Sperber - 7 years, 1 month ago

Thats how I did it. Simple and pretty.

Yamna Khan - 7 years, 1 month ago
Frank Rodriguez
Apr 27, 2014

you can use pascals triangle to solve this and examine the middle term divided by the sum. For 2, its 1 2 1. 2/4=.5; For 4, its 1 4 6 4 1. 6/16=.375; For 6, its 1 6 15 20 15 6 1. 20/64= .3125. So by this logic, f(200) is less than f(100).

Kawaii Kuma
Apr 27, 2014

Why everyone are so complicated T_T. Hospital A has 1/100 chance of having equally boys and girls while Hospital B has 1/200 chance :. It's very easy :\

How did you get that formula??... I don't understand :/

puzzle seeker - 7 years, 1 month ago

Because this question is not as easy as you think. :/ :\

Samuraiwarm Tsunayoshi - 6 years, 9 months ago
Dhritiman Dutta
Apr 19, 2014

Probability tends to more true when the sample space large so , hospital B has greater sample space than hospital A, So hospital B has greater chance.

That is incorrect. It is more likely to be CLOSE to even, but the probability of exactly even goes down. It's a simple binomial distribution problem.

Ron Sperber - 7 years, 1 month ago

Why we might think that bigger sample is better for getting the same number of boys and girls: we simply expect that all the variations cancel out, but in this specific task, those variations have to cancel out completely, which makes it less probable.

Still, it is true that if we plot the distribution of boys and girls (for every day of the year), we expect that hospital with greater number of kids should have smaller relative difference between boys and girls throughout the year.

Alija Bevrnja - 7 years, 1 month ago
Cindy Wu
May 20, 2014

Funny how I, without complex A-level maths, managed to solve this using intuition only. I felt that the odds are, realistically, never even. So, the less the number of total births, the less likely it is that there are outstanding uneven numbers of boys and girls. Therefore it is A. Given what was just said though, I did ignore the bit that said the chances were exactly even, because that gave two options as an answer, which of course means by my logic one should probably skip over that detail somewhat

Henrique Ribeiro
Apr 24, 2014

As hospital B has more babies, the chances to have equal number of boys and girls are hardest.

Vinit Shandilya
Apr 24, 2014

Probability p = favourable cases/Total no. of cases. Here, favourable case is just one, 50% boys and 50% girls. Total cases include: 400 boys: 0 girls, 399 boys: 1 girl.. and so on. So total number of cases are 400 and 200. Therefore, pA = 1/200, pB=1/400

Vibhu Baibhav
Apr 24, 2014

its nothing maths in it just use ur iq. out of 200 babies probability will be more of equal no. of babies than in 400 if u have 2 balls and someone else has 5 balls have more probability of having the balls of same color

Guilherme Vargas
Apr 24, 2014

If we have 400 babies, there are 401 possible combinations of babies:

--- 0 boy, 400 girls

--- 1 boy, 399 girls

--- 2 boys, 398 girls

..........

--- 400 boys, 0 girls

And in only one case, the numbers are equal(200,200). So, the chance of it happening, for 400 babies, is 1/401.

If the same thinking is applied to the other hospital, we come to a probability of 1/201.

As 1/201>1/401, probability of hospital A is bigger than hospital B.

This is why i think the probability field needs a new Gauss. We tend to over simplify very complex problems and take answers as right that we know are wrong in our gut just because we apply some current formulas and accept the answers without a critical spirit. We are all saying that there is exactly the same probability of having 200 girls + 200 boys or 400 girls and 0 boys which seems right from our blind formula usage but we all know in our gut that this isn't true. There is only one route to 400G + 0B but there are a lot of ways to get to 200G + 200B. Just to end my point, if for 100 years i own a hospital that every year delivers 400 babies i can say that in some of those years i will certainly get 200G + 200B and can also say that will never get a year with 400G + 0B. We all know this is true and we keep saying the probability is the same for both cases.

Miguel Rocha - 7 years, 1 month ago

In hospital B of 400 babies, probability will be half that of hospital A of 200 babies. Thus Hospital A has more chance!

Fredrik Meyer
Apr 19, 2014

The number of ways to produce n of 2n boys is ( 2 n n ) \binom{2n}{n} . By Stirling's formula, this is approximately 4 n π n \frac{4^n}{\sqrt{\pi n}} . Thus the probability of having exactly half boys half girls is 4 n π n / 2 2 n = 1 π n \frac{4^n}{\sqrt{\pi n}} / 2^{2n} = \frac{1}{\sqrt{ \pi n}} . Thus the greater the number of trials, the more difficult it is to get exactly half.

Samir Shaheen
Apr 19, 2014

It is simple as the population size increases the chances if reaching the universal average increases. That is why hospital A with small number of babies has a higher chance of diverting from the universal average and having equal number if boys and girls.

But the task is very specific. Let's say we flip a coin 1 million times today and then tomorrow we flip it only twice. On which day we have greater chance of getting the same number of heads and tails?

Alija Bevrnja - 7 years, 1 month ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...