1990 Paper 1 Q15

Year: 1990
Paper: 1
Question Number: 15

Course: LFM Stats And Pure
Section: Negative Binomial Distribution

Difficulty: 1500.0 Banger: 1591.4

Problem

A coin has probability \(p\) (\(0 < p < 1\)) of showing a head when tossed. Give a careful argument to show that the \(k\)th head in a series of consecutive tosses is achieved after exactly \(n\) tosses with probability \[ \binom{n-1}{k-1}p^{k}(1-p)^{n-k}\qquad(n\geqslant k). \] Given that it took an even number of tosses to achieve exactly \(k-1\) heads, find the probability that exactly \(k\) heads are achieved after an even number of tosses. If this coin is tossed until exactly 3 heads are obtained, what is the probability that exactly 2 of the heads occur on even-numbered tosses?

Solution

We must have a sequence consisting of \(\underbrace{HTT\cdots TH}_{k-1\text{ heads and }n-k\text{ tails}}\underbrace{H}_{k\text{th head}}\). There are \(\binom{n-1}{k-1}\) ways to chose how to place the \(k-1\) heads in the first \(n-1\) flips, and each sequence has probability \(p^{k-1}(1-p)^{n-k}p\) which gives a probability of \(\displaystyle \binom{n-1}{k-1} p^k (1-p)^{n-k}\). Given that it took an even number of tosses to achieve \(k-1\) heads, this is equivalent to the problem of what is the probability that the first head occurs on an even flip, ie \begin{align*} \mathbb{P}(\text{even flip}) &= \mathbb{P}(2\text{nd flip}) +\mathbb{P}(4\text{th flip}) +\mathbb{P}(6\text{th flip}) + \cdots \\ &= (1-p)p + (1-p)^3p + (1-p)^5p + \cdots \\ &= (1-p)p \left ( \sum_{r=0}^\infty (1-p)^{2r}\right) \\ &= \frac{p(1-p)}{1-(1-p)^2} \\ &= \frac{p(1-p)}{2p-p^2} \\ &= \frac{1-p}{2-p} \end{align*} The ways to achieve \(2\) heads on even tosses are \(EEO\), \(EOE\), \(OEE\). The probability of going from \(O\) to \(E\) is the same as the initial probability of an \(O\) flip, etc. Therefore \begin{align*} \mathbb{P}(EEO) &=\left( \frac{1-p}{2-p} \right)^2 \left ( 1- \frac{1-p}{2-p} \right) \\ &= \left( \frac{1-p}{2-p} \right)^2 \left ( \frac{1}{2-p} \right) \\ \mathbb{P}(EOE) &= \left( \frac{1-p}{2-p} \right) \left ( \frac{1}{2-p} \right)^2 \\ \mathbb{P}(OEE) &= \left ( \frac{1}{2-p} \right)^2 \left( \frac{1-p}{2-p} \right)\\ \mathbb{P}(2 \text{ heads on even tosses}) &= \frac{(1-p)^2 + 2(1-p)}{(2-p)^3} \\ &= \frac{(1-p)(2-p)}{(2-p)^3} \\ &= \frac{1-p}{(2-p)^2} \end{align*}
Rating Information

Difficulty Rating: 1500.0

Difficulty Comparisons: 0

Banger Rating: 1591.4

Banger Comparisons: 8

Show LaTeX source
Problem source
A coin has probability $p$ ($0 < p < 1$) of showing a head when tossed. Give a careful argument to show that the $k$th head in a series of consecutive tosses is achieved after \textit{exactly} $n$ tosses with probability 
\[
\binom{n-1}{k-1}p^{k}(1-p)^{n-k}\qquad(n\geqslant k).
\]
Given that it took an even number of tosses to achieve exactly $k-1$ heads, find the probability that exactly $k$ heads are achieved after an even number of tosses. 
If this coin is tossed until exactly 3 heads are obtained, what is the probability that \textit{exactly }2 of the heads occur on even-numbered
tosses?
Solution source
We must have a sequence consisting of $\underbrace{HTT\cdots TH}_{k-1\text{ heads and }n-k\text{ tails}}\underbrace{H}_{k\text{th head}}$. There are $\binom{n-1}{k-1}$ ways to chose how to place the $k-1$ heads in the first $n-1$ flips, and each sequence has probability $p^{k-1}(1-p)^{n-k}p$ which gives a probability of $\displaystyle \binom{n-1}{k-1} p^k (1-p)^{n-k}$.

Given that it took an even number of tosses to achieve $k-1$ heads, this is equivalent to the problem of what is the probability that the first head occurs on an even flip, ie

\begin{align*}
\mathbb{P}(\text{even flip}) &= \mathbb{P}(2\text{nd flip}) +\mathbb{P}(4\text{th flip}) +\mathbb{P}(6\text{th flip}) + \cdots \\
&= (1-p)p + (1-p)^3p + (1-p)^5p + \cdots \\
&= (1-p)p \left ( \sum_{r=0}^\infty (1-p)^{2r}\right) \\
&= \frac{p(1-p)}{1-(1-p)^2} \\
&= \frac{p(1-p)}{2p-p^2} \\
&= \frac{1-p}{2-p}
\end{align*}

The ways to achieve $2$ heads on even tosses are $EEO$, $EOE$, $OEE$. The probability of going from $O$ to $E$ is the same as the initial probability of an $O$ flip, etc.

Therefore

\begin{align*}
\mathbb{P}(EEO) &=\left( \frac{1-p}{2-p} \right)^2 \left ( 1- \frac{1-p}{2-p} \right) \\
&=  \left( \frac{1-p}{2-p} \right)^2 \left ( \frac{1}{2-p} \right) \\
\mathbb{P}(EOE) &= \left( \frac{1-p}{2-p} \right) \left ( \frac{1}{2-p} \right)^2 \\
\mathbb{P}(OEE) &=  \left ( \frac{1}{2-p} \right)^2 \left( \frac{1-p}{2-p} \right)\\
\mathbb{P}(2 \text{ heads on even tosses}) &= \frac{(1-p)^2 + 2(1-p)}{(2-p)^3} \\
&= \frac{(1-p)(2-p)}{(2-p)^3} \\
&= \frac{1-p}{(2-p)^2}
\end{align*}