Is it really true that statistics is useless, since you can design an experiment to statistically prove anything you want to prove?

False

While its true that you can design an experiment to statistically prove just about anything you want to prove, such an experiment would have to be very poorly designed, and statistically flawed. The purpose of statistics is to ensure that incorrect conclusions are NOT drawn, not to mislead people by proving that incorrect conclusions are true. A well designed study, backed by solid statistics, will usually give reliable results.