= Quadratic Galton-Watson extinction bound
{title2=$q\leq1-2(\mu-1)/(\sigma^2+\mu^2-\mu)$}
Let a <Galton-Watson process> have offspring count $K$ with finite <variance> $\sigma^2$ and <expected value> $\mu>1$. Its <branching extinction probability> $q$ satisfies
$$
q\leq1-\frac{2(\mu-1)}{\mathbb E[K(K-1)]}
=1-\frac{2(\mu-1)}{\sigma^2+\mu^2-\mu}.
$$
To prove the bound, write $f$ for the offspring <probability generating function> and $r=1-q>0$ for the <branching survival probability>. Nonnegative coefficients give $f''(s)\leq f''(1)$ on $[0,1]$. The integral remainder in <Taylor's theorem> therefore gives
$$
f(1-r)\leq1-\mu r+\tfrac12f''(1)r^2.
$$
The <Galton-Watson extinction fixed point> says $f(1-r)=1-r$. Cancelling and dividing by $r$ gives $r\geq2(\mu-1)/f''(1)$, with $f''(1)=\mathbb E[K(K-1)]$. Thus two offspring moments already give a quantitative survival guarantee, without solving the full fixed-point equation.
For $P(K=0)=1/4$ and $P(K=2)=3/4$, the <expected value> is $3/2$ and $\mathbb E[K(K-1)]=3/2$, so the bound gives $q\leq1/3$. The offspring <probability generating function> is $f(s)=1/4+3s^2/4$, whose smallest fixed point is exactly $1/3$; this example attains the bound.
Back to article page