\[
\newcommand{\C}{\mathbb{C}}
\newcommand{\haar}{\mathsf{m}}
\newcommand{\P}{\mathcal{P}}
\newcommand{\R}{\mathbb{R}}
\newcommand{\Q}{\mathbb{Q}}
\newcommand{\N}{\mathbb{N}}
\newcommand{\Z}{\mathbb{Z}}
\newcommand{\d}{\mathsf{d}}
\newcommand{\g}{>}
\newcommand{\l}{<}
\newcommand{\intd}{\,\mathsf{d}}
\newcommand{\Re}{\mathsf{Re}}
\newcommand{\area}{\mathop{\mathsf{Area}}}
\newcommand{\met}{\mathop{\mathsf{d}}}
\newcommand{\emptyset}{\varnothing}
\DeclareMathOperator{\borel}{\mathsf{Bor}}
\]
Week 7 Tutorial Solutions
- $\mu_p([01]) = (1-p)p$
- $\mu_p([01]) = (1-p)p$ because $1_{[01]}$ is simple.
- $Z_1 = 1_{[11]}$ is simple so its integral is $\mu_p([11]) = p^2$.
- $T^{-1}([01]) = [001] \cup [101]$ is a disjoint union so
\[
\mu_p(T^{-1}([01])) = \mu_p([001]) \cup \mu_p([101])
\]
and $\mu_p(T^{-1}([01])) = (1-p)^2 p + p(1-p)p = (1-p)p$.
- $Z_2 = 1_{T^{-1}[11]}$ is simple. Its integral is $\mu_p(T^{-1}[11])$ which is
\[
\mu_p([011]) + \mu_p([111]) = p^2
\]
as $[011]$ and $[111]$ are disjoint.
- Calculate
\[
\begin{align*}
\int Y_n \intd \mu_p &{} = \int Z_n \left( - \int Z_n \intd \mu_p \right) \intd \mu_p \\
&{} = \int Z_n \intd \mu_p - \int Z_n \intd \mu_p = 0
\end{align*}
\]
by linearity and the fact that $\mu_p(X) = 1$.
- From
\[
\begin{align*}
Y_n Y_{n+1} &{} = Z_n Z_{n+1} - p^2 Z_n - p^2 Z_{n+1} + p^4 \\
&{} = X_n X_{n+1} X_{n+2} - p^2 X_n X_{n+1} - p^2 X_{n+1} X_{n+2} + p^4
\end{align*}
\]
we can calculate
\[
\int Y_n Y_{n+1} \intd \mu_p = p^3 - p^4
\]
- From
\[
\begin{align*}
Y_n Y_{n+2} &{} = Z_n Z_{n+2} - p^2 Z_n - p^2 Z_{n+2} + p^4 \\
&{} = X_n X_{n+1} X_{n+2} X_{n+3} - p^2 X_n X_{n+1} - p^2 X_{n+2} X_{n+3} + p^4
\end{align*}
\]
we can calculate
\[
\int Y_n Y_{n+2} \intd \mu_p = 0
\]
- The integral
\[
\begin{align*}
& Y_a Y_b Y_c Y_d \\
= {}& (X_a X_{a+1} - p^2)(X_b X_{b+1} - p^2)(X_c X_{c+1} - p^2)(X_d X_{d+1} - p^2)
\end{align*}
\]
will be zero if
\[
a \notin \{ b-1,b,b+1,c-1,c,c+1,d-1,d,d+1 \}
\]
or the same holds for any permutation of the indices.
- There is a constant $E \g 0$ such that at most $EN^2$ of all possible choices $1 \le a,b,c,d \le N$ result in a non-zero integral in the previous question. It is in any case bounded by 1 so our proof from the notes of the strong law of large numbers works in this setting as well.