# Can DQN perform better than Double DQN?

I'm training both kind of agents against an environment but DQN performs significantly better than Double DQN. As I've saw here, Double DQN use to perform better than DQN. Am I doing something wrong or is it possible?