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

This interesting question was asked at a while ago but got no answers. The original author is Jesko Hüttenhain.

Consider the Chernoff bound described in Theorem 1 of this paper:

Theorem 1. Let $X_1,\ldots,X_n$ be discrete, independent random variables such that $E[X_i] = 0$ and $|X_i|<1$ for all $i$. Let $X:=\sum_{i=1}^n X_i$ and $\sigma^2$ be the variance of $X$. Then, $$\Pr\left[|X|\ge \lambda\sigma\right] \le 2e^{-\lambda^2/4}$$ for any $0\le\lambda\le 2\sigma$.

I want to apply this estimator for a computation, but the variance of the variables $X_i$ is unknown to me. Apart from that, my variables satisfy all the conditions of Theorem 1. In fact, my variables are independent and identically distributed (iid).

On the other hand, there is Chebyshev's inequality for finite samples which does not require knowledge of the variance (or mean) of a given random variable and replaces it with the corresponding values of my given sample. However, the estimator is not good enough for my application and I was wondering if there is an estimator similar to Theorem 1 but not featuring the variance of the distribution itself.

Intuitively speaking, it would be nice to have the best of both worlds: If that is not possible, then what is the best bound I can achieve for a sum of iid variables, without any knowledge about their variance?

share|cite|improve this question
There's at least one mistake in the ``paper'' you refer to. Neither side depends on $n$ but for every $K$ we must have $\mathbb P\left[X > K\right] \to 1$ by the central limit theorem. If there's only one mistake the exponent in the right hand side should be $\left(2n-1\right)\frac{\lambda^2}4$ but go through it very carefully if you want to use this. Also if $|X_1|<1$ and $\mathbb E(X) = 0$ you must have $\sigma^2<1$ by Chebychev's inequality so know the variance. – user32372 Apr 18 '13 at 8:13
Thank you for spotting that mistake. – tjeng Apr 18 '13 at 11:20

I would use Hoeffding's inequality. It exploits the boundedness of the random variables (like your Theorem 1), only without the variance making an appearance, and you still get exponential decay in the probability estimate (unlike Chebyshev's inequality).

For a big-picture perspective on the different types of concentration inequalities available, check out Terry Tao's notes on the subject.

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.