Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $X_0$,$X_1$, a Galton-Watson branching process. Let us denote $\epsilon$ for the probability when $X_0 = 1$ that the population eventually becomes extinct (that is, that $X_n = 0$ for all suciently large $n$). Then, Set $Z_n = \epsilon^{X_n}$ for $n\geq 0$.

I'm trying to show that $Z_0$,$Z_1$ a martingale.

My proof:

$E[Z_n|Z_{n-1}] = E[\epsilon^{X_n}|Z_{n-1}]$,but not sure how to carry this out further.

share|cite|improve this question
up vote 1 down vote accepted

This uses some nice properties of $\epsilon$. Let $Y_{ni}$ be iid and take on values in $\{0,1,2,\ldots\}$ with respective probabilities $\{p_0,p_1,\ldots\}$. Then the branching process looks something like $X_n=\sum_{i=1}^{X_{n-1}}Y_{ni}$ where by convention the sum is zero if $X_{n-1}=0$.

Conditioning on $Z_{n-1}$ effectively determines $X_{n-1}$, so that

$\mathbb{E}[Z_n|Z_{n-1}]= \mathbb{E}[\epsilon^{\sum_{i=1}^{X_{n-1}}Y_{ni}}|Z_{n-1}]=\mathbb{E}[\epsilon^{Y_{n1}}]^{X_{n-1}}$

where I used independence of $Y_{ni}$ and that $X_{n-1}$ is determined by $Z_{n-1}$ to get rid of conditioning. So we need to show $\mathbb{E}[\epsilon^{Y_{n1}}]=\epsilon$ for this to be a martingale. Notice that we can explicitely write:

$\mathbb{E}[\epsilon^{Y_{n1}}]=\sum_{k=0}^\infty \epsilon^kp_k$

Call $D$ the event that extinction occurs. Notice that for $D$ to occur every single subtree has to die for each first generation member. So we should look at what happens for $X_1=k$ for all $k$. If $k=0$ then extinction has occured. We write,

$\epsilon:=P(D|X_0=1)\\ =\sum_{k=0}^\infty P(D\cap (X_1=k)|X_0=1)\\ =\sum_{k=0}^\infty P(D|X_1=k \cap X_0=1)P(X_1=k|X_0=1)\\ =\sum_{k=0}^\infty \epsilon^kp_k$

where we used a number of facts: $P(A\cap B|C) = P(A| B\cap C)P(B|C)$ and the fact that extinction probability at the first level is the same as the zeroth level. More precisely, if we have $X_1\geq 1$, then the Galton Watson process started on the $X_1$ subtree is independent of what happened in $X_0$. In short, the Galton Watson process is Markov in that conditioning on the past reduces to just what the previous generation was at.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.