Serving the Quantitative Finance Community

 
User avatar
modraig
Topic Author
Posts: 0
Joined: January 18th, 2010, 7:18 am

Newborns brainteaser

February 20th, 2010, 6:33 pm

Hi all,Got the following question in an interview and due to the different answers i got from friends I would like your opinion too.In a hospital room for newborns, there are 2 boys and 3 girls. A woman gives birth to a baby and this is also added to the room. we randomly select 1 baby and it is a boy. What is the probability that the baby is a boy?Thanks in advance.
 
User avatar
wileysw
Posts: 7
Joined: December 9th, 2006, 6:13 pm

Newborns brainteaser

February 20th, 2010, 7:09 pm

say the gender ratio at birth is 1:1 for boy and girl (more realistic number could be found here), then the prob that the new-born is a boy is (3/6)/(3/6+2/6)=0.6if you search on the forum, there are couple of old threads discussing similar problems (e.g., here)----- ----- ----- ----- -----not sure if this is the original source (problem 56 "babies in the nursery problem"), but as noticed by avishalom below, the girl # is irrelevant for the answer.
Last edited by wileysw on February 22nd, 2010, 11:00 pm, edited 1 time in total.
 
User avatar
modraig
Topic Author
Posts: 0
Joined: January 18th, 2010, 7:18 am

Newborns brainteaser

February 20th, 2010, 10:07 pm

Hi Wileyw,and thank you for your quick reply! I have read the indicated thread and I would like to ask you if it is possible to tell me how you came up with this solution. The tree we have created is described as follows:Boys (2/6)Girls (3/6)Random (1/6) which continues to another tree with total pr(boy)=(1/6)*(1/2) and same for Pr(girl)=(1/6)*(1/2)Then calculated Pr(random|boy)=Pr(random&boy)/Pr(boy)=[(1/6)*(1/2)]/[2/6+(1/6)*(1/2)]=1/5Based on your answer, I believe you are using a different approach. I would really appreciate if you could tell me how you did it.Thanks
 
User avatar
Eddi
Posts: 0
Joined: February 9th, 2010, 4:45 am

Newborns brainteaser

February 20th, 2010, 10:22 pm

What wiley has reads: Pr(newborn is a boy|selected boy) = Pr(newborn is a boy & selected boy)/Pr(selected boy) = Pr(newborn is a boy & selected boy)/ (Pr(selected boy|newborn is a boy) + Pr(selected boy|newborn is a girl))
 
User avatar
modraig
Topic Author
Posts: 0
Joined: January 18th, 2010, 7:18 am

Newborns brainteaser

February 21st, 2010, 4:34 am

Hi Eddi,2 questions1) Shouldnt the denominator be (Pr(selected boy&newborn is a boy) + Pr(selected boy&newborn is a girl))? Otherwise you have to use something like this: P(B) = P(A&B) + P(A'&B) = P(B|A) P(A) + P(B|A')P(A') (ref. wikipedia)2) Could you please explain to me why the Pr(newborn is a boy & selected boy)=3/6? The random baby has a 1/2 probability to be a boy not 1. Sorry for asking all these questions just trying to understand how it works. maybe the tree i have in my mind is incomplete.Thanks for all the answers.
Last edited by modraig on February 20th, 2010, 11:00 pm, edited 1 time in total.
 
User avatar
wileysw
Posts: 7
Joined: December 9th, 2006, 6:13 pm

Newborns brainteaser

February 21st, 2010, 6:34 am

modraig, sorry about the sloppy notation. essentially this is a common application of Bayes' theorem (see e.g., here)say A = newborn is a boy, B = random selected baby (from 6) is a boy.as Eddi indicated, we want P(A|B), i.e., the posterior prob of A given B observed.Bayes' theorem: P(A|B) = P(B|A)*P(A)/P(B)P(A)=1/2 by assumption;if the newborn is a boy, there is P(B|A)=3/6 chance of selecting a boy.now P(B)=P(B|A)*P(A)+P(B|A')*P(A') as you said, where A' = newborn is a girl.similarly, P(A')=1/2 by assumption, and P(B|A')=2/6.hence the answer is (1/2)*(3/6) / [(1/2)*(3/6)+(1/2)*(2/6)] = 0.6(my previous post did not include these 1/2 factors)----- ----- ----- ----- -----your expression seems to answer the following question:given the randomly selected baby is a boy, what is the prob of him being the newborn one?
Last edited by wileysw on February 20th, 2010, 11:00 pm, edited 1 time in total.
 
User avatar
avishalom
Posts: 0
Joined: October 8th, 2009, 8:26 am

Newborns brainteaser

February 22nd, 2010, 9:02 am

Cool so the original number of girls in the room does not matter at all. it cancels out. so if you originally had 0 girls or 1miliion girls doesn't matter.
 
User avatar
antiser
Posts: 0
Joined: June 18th, 2009, 12:25 pm

Newborns brainteaser

February 22nd, 2010, 7:02 pm

Wow! that is a little counter-intuitive!
 
User avatar
karakfa
Posts: 0
Joined: May 25th, 2002, 5:05 pm

Newborns brainteaser

February 22nd, 2010, 9:27 pm

yep, the simplified formula is:p = P(boy born | boy selected) = (b + 1) / (2b + 1), where b is the number of boys initially.some special cases:b=0 -> p=1b->Inf -> p=0.5
 
User avatar
marvas
Posts: 0
Joined: February 26th, 2010, 10:26 am

Newborns brainteaser

February 26th, 2010, 10:54 am

Guys if you think rationally your answer must be wrong.If you make the probability tree and assume 50% chance of new born is boy or girl ( i.e. independant of the current population)then the probability you will get a boy randomly from the room cannot be more than 50%.Becuase a priori we have 2 boys and 3 girls making the probability of a boy be 2/5.After adjusting for the new born, in all possible configurations, the probability should be 5/12.Are you asking the right question? This to me is wrong.The tree is like thisNew born is boy (1/2) P(boy given new born is boy) 3/6 P(girl ..) 3/6NewBord is girl(1/2) P(boy ... is girl) 2/6 P(girl ...) 4/6given this space the total probability of boy is 5/6 / (4+3+5)/6
Last edited by marvas on February 25th, 2010, 11:00 pm, edited 1 time in total.
 
User avatar
marvas
Posts: 0
Joined: February 26th, 2010, 10:26 am

Newborns brainteaser

February 26th, 2010, 10:57 am

Can someone restate the question. Are we asking for the probability the baby born is a boy or probability of a random child from that ward being a boy?
 
User avatar
tristanm
Posts: 0
Joined: October 20th, 2009, 3:21 pm

Newborns brainteaser

February 26th, 2010, 12:00 pm

QuoteOriginally posted by: marvasCan someone restate the question. Are we asking for the probability the baby born is a boy or probability of a random child from that ward being a boy?The probability that a random child picked from the ward is a boy, given that there are 2 boys, 3 girls and one unknown.
 
User avatar
marvas
Posts: 0
Joined: February 26th, 2010, 10:26 am

Newborns brainteaser

February 28th, 2010, 2:57 pm

so i am right prob should be 5/12
 
User avatar
WutsaQuant
Posts: 0
Joined: November 2nd, 2009, 9:46 pm

Newborns brainteaser

February 28th, 2010, 11:50 pm

QuoteOriginally posted by: marvasCan someone restate the question. Are we asking for the probability the baby born is a boy or probability of a random child from that ward being a boy?What is the probability that the baby born was a boy, given that a random child from that ward is a boy.In this context, the answer is 3/5 as mentioned earlier. edit: grammar
Last edited by WutsaQuant on February 28th, 2010, 11:00 pm, edited 1 time in total.
 
User avatar
FritzJacob
Posts: 3
Joined: January 22nd, 2010, 11:15 am

Newborns brainteaser

March 4th, 2010, 2:27 am

Last edited by FritzJacob on March 3rd, 2010, 11:00 pm, edited 1 time in total.