Main Page | Report this Page
 
   
Science Forum Index  »  Statistics - Math Forum  »  Question on counting process
Page 1 of 1    
Author Message
TheTravellingSalesman
Posted: Sat Apr 05, 2008 3:56 pm
Guest
Can anyone help me with this question?

At the end of an election, the eventual winner, candidate B, receives
m votes, and
H receives n < m votes. Assuming that as the votes are counted they
come as
independent random variables with probability half that they are for
either candidate,
show that the probability B always stayed ahead of H through the
counting process
is given by
(m - n)/(m + n)
Jack Tomsky
Posted: Sun Apr 06, 2008 5:02 pm
Guest
Quote:
On Sat, 5 Apr 2008 18:56:10 -0700 (PDT),
TheTravellingSalesman
saad.zaman@gmail.com> wrote:

Can anyone help me with this question?

At the end of an election, the eventual winner,
candidate B, receives
m votes, and
H receives n < m votes. Assuming that as the votes
are counted they
come as
independent random variables with probability half
that they are for
either candidate,
show that the probability B always stayed ahead of
H through the
counting process
is given by
(m - n)/(m + n)

I've never seen that problem, and I'm surprised that
the answer is apparently that simple.

Well, it is true for n+m = 3. Maybe you can show it
by induction?

--
Rich Ulrich

http://www.pitt.edu/~wpilib/index.html



Rich, this is known as the ballot problem. It counts the number of paths from (0,0) to (m+n,m-n) for which the y coordinate is positive from x=1 on, divided by the total number of paths. This gives the probability that B is always ahead of H during the entire counting process. You can find the derivation in Feller and similar books.

Jack
Richard Ulrich
Posted: Sun Apr 06, 2008 9:12 pm
Guest
On Sat, 5 Apr 2008 18:56:10 -0700 (PDT), TheTravellingSalesman
<saad.zaman@gmail.com> wrote:

Quote:
Can anyone help me with this question?

At the end of an election, the eventual winner, candidate B, receives
m votes, and
H receives n < m votes. Assuming that as the votes are counted they
come as
independent random variables with probability half that they are for
either candidate,
show that the probability B always stayed ahead of H through the
counting process
is given by
(m - n)/(m + n)

I've never seen that problem, and I'm surprised that
the answer is apparently that simple.

Well, it is true for n+m = 3. Maybe you can show it
by induction?

--
Rich Ulrich

http://www.pitt.edu/~wpilib/index.html
danheyman@yahoo.com
Posted: Mon Apr 07, 2008 11:32 am
Guest
On Apr 5, 9:56 pm, TheTravellingSalesman <saad.za...@gmail.com> wrote:
Quote:
Can anyone help me with this question?

At the end of an election, the eventual winner, candidate B, receives
m votes, and
H receives n < m votes. Assuming that as the votes are counted they
come as
independent random variables with probability half that they are for
either candidate,
show that the probability B always stayed ahead of H through the
counting process
is given by
(m - n)/(m + n)

Look up "ballot theorems" in a reference book or online.
Richard Ulrich
Posted: Mon Apr 07, 2008 10:29 pm
Guest
On Sun, 06 Apr 2008 23:02:47 EDT, Jack Tomsky <jtomsky@ix.netcom.com>
wrote:

Quote:
On Sat, 5 Apr 2008 18:56:10 -0700 (PDT),
TheTravellingSalesman
saad.zaman@gmail.com> wrote:

Can anyone help me with this question?

At the end of an election, the eventual winner,
candidate B, receives
m votes, and
H receives n < m votes. Assuming that as the votes
are counted they
come as
independent random variables with probability half
that they are for
either candidate,
show that the probability B always stayed ahead of
H through the
counting process
is given by
(m - n)/(m + n)

I've never seen that problem, and I'm surprised that
the answer is apparently that simple.

Well, it is true for n+m = 3. Maybe you can show it
by induction?

--
Rich Ulrich

http://www.pitt.edu/~wpilib/index.html



Rich, this is known as the ballot problem. It counts the number of paths from (0,0) to (m+n,m-n) for which the y coordinate is positive from x=1 on, divided by the total number of paths. This gives the probability that B is always ahead of H during the entire counting process. You can find the derivation in Feller and similar books.

Jack,
Thanks. Even with a name given to it,
I don't remember seeing it.

--
Rich Ulrich

http://www.pitt.edu/~wpilib/index.html
 
Page 1 of 1       All times are GMT - 5 Hours
The time now is Tue Oct 07, 2008 9:25 am