DDPG++: Striving for Simplicity in Continuous-control Off-Policy Reinforcement Learning