David Williams "Probability with Martingales": Martingales in $L^2$ - orthogonality of increments

513 Views Asked by At

Consider this extract from the above book:

orthog1 orthog2

My question: should the first highlight 12.1.b not be $$ E[{(M_n-M_0)}^2] = \sum_1^n E[{(M_k-M_{k-1})}^2] $$

It would seem more compatible with the second highlight 12.1.d and formula 12.1.a

Many thanks in advance !

1

There are 1 best solutions below

0
On BEST ANSWER

Using the tower property and the martingale property, we find

$$\mathbb{E}(M_0 M_n) = \mathbb{E} \big( \mathbb{E}(M_0 M_n \mid \mathcal{F}_0) \big) = \mathbb{E}\big( M_0 \underbrace{\mathbb{E}(M_n \mid \mathcal{F}_0)}_{M_0} \big) = \mathbb{E}(M_0^2)$$

and therefore

$$\mathbb{E}((M_n-M_0)^2) = \mathbb{E}(M_n^2)-2 \mathbb{E}(M_n M_0) + \mathbb{E}(M_0^2) = \mathbb{E}(M_n^2) - \mathbb{E}(M_0^2).$$

This shows that

$$\mathbb{E}(M_n^2) = \mathbb{E}(M_0^2) + \sum_{k=1}^n \mathbb{E}((M_k-M_{k-1})^2)$$

is equivalent to saying

$$\mathbb{E}((M_n-M_0)^2) = \sum_{k=1}^n \mathbb{E}((M_k-M_{k-1})^2)$$