Probability of error in a communications channel is measured experimentally and determined using the relative frequency definition: (error) = / where is the total number of bits sent over the channel, and is the number of bits received in error. Assuming that each bit transmission is independent of other transmissions, how many bits do we have to send in order to be 99.9% certain that the observed relative frequency is within 5% of the true probability of error, which is assumed to be less than 0.01