It's so crazy to me that they wouldn't have DNA with such a quick and vicious crime. The person who did this knew he could be discovered at any time so wouldn't have lingered.
JMO, but there's a difference between "we found zero DNA from the perpetrator at this crime scene" and "there was DNA but it was present in a mix of other contributors to the extent that it is not amenable to analysis under current techniques."
Because in the first scenario you find yourself asking "how did this mastermind criminal eliminate all traces of his DNA from the scene?" Which seems almost impossible. But in the second scenario you realize that he could just be very lucky. Perhaps a victim had his DNA under her fingernails but the DNA of the victim's sibling and even another person was also present. Perhaps some of his touch DNA was on a sweatshirt of one of the victims but that sweatshirt had previously been worn by many high school friends whose DNA was on it too. The power of many probabilistic genotyping programs decreases greatly when either 1. There are more than three contributors or 2. The analyst cannot make a good assumption about how many contributors there were. In my opinion only.