# Hoeffding's inequality within not a range of values but only two possible values

We have to create confidence intervals for a certain probability p of n i.i.d Bernoulli random variables (Xn) that can take only two specific values {red,blue}. The confidence interval must be exponentially high.

The thing is, I'm positive you can approach the problem with Hoeffeding's inequality, however I'm not sure how to plug in those "red",blue" values.

Any ideas?

-
Got it, there's a special case for Hoeffding's inequality for Bernoulli variables. That's enough. –  Rotational Nov 13 '12 at 0:16