"Too small" means either it can not be detected experimentally, or even in principle, depending on whether you are allowing gedankenexperiments.

I don't think it is reasonable to expect our models to be identical to reality. Otherwise, they wouldn't be models, they would be reality. A theory is correct if it produces experimentally verified predictions.