MathOverflow is a question and answer site for professional mathematicians. 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

Hi everyone, please consider the following problem:

Let $(M_t)_{t\geq 0}$ be a continuous and positive submartingale and $S_t=\sup_{0\leq s\leq t}M_s$. Please prove that for any $\lambda>0$ we have

$$\lambda P(S_t>2\lambda)\leq E[M_t1_{\{M_t>\lambda\}}]$$

This inequality makes me remember the Doob inequality

$$\lambda P(S_t>\lambda)\leq E[M_t1_{\{S_t>\lambda\}}]$$

So it is enough to show that

$$E[M_t1_{\{S_t>\lambda\}}]\leq 2E[M_t1_{\{M_t>\lambda\}}]$$

But I have no idea to deal about the term $1_{\{M_t>\lambda\}}$, even by introducing a stopping time $T_{\lambda}$. Could someone help me prove this inequality or give some idea? Many thanks!

share|cite|improve this question
There's a typo in your title: Dood -> Doob. – HJRW Oct 24 '12 at 20:38
How do you know that this inequality holds? – Alekk Oct 24 '12 at 21:59
up vote 1 down vote accepted

Nice little exercise, it should go to, though. I'll give you a hint: Let $T$ be the hitting time of $2\lambda$. Then, $E[M_t 1_{M_t\ge \lambda}] \ge E[M_t 1_{M_t\ge \lambda,\ T\le t}] \ge E[M_t1_{T\le t}]- E[M_t1_{M_t\le \lambda}1_{T\le t}].$

share|cite|improve this answer
Thanks so much for the hint of Pascal Maillard. Here is a proof: By Doob inequality: $$E[M_t1_{\{T\leq t\}}]\geq 2\lambda P(S_t>2\lambda)$$ and $$E[M_t1_{\{M_t\leq\lambda, T\leq t\}}]\leq E[\lambda 1_{\{M_t\leq\lambda, T\leq t\}}]\leq\lambda P(S_t>\lambda)$$ which gives $$E[M_t1_{\{T\leq t\}}]-E[M_t1_{\{M_t\leq\lambda, T\leq t\}}]\geq 2\lambda P(S_t>\lambda)-\lambda P(S_t>\lambda)=\lambda P(S_t>\lambda)$$ That is what we need! – Higgs88 Oct 25 '12 at 13:32

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.