Velocity verlet algorithm

I've read that this algorithm conserves energy if the system it's applied to conserves energy. I can't find a proof, and it's not a particularly obvious statement, so how would you prove it?
Actually, Verlet integration is approximately conserving the (total) energy, it does however preserve total angular momentum. This is the reason it is used for orbital mechanics problems. The properties of the propagation matrix of the discretized system can already give you a lot of information, especially if you do it for the hamiltonian. There is a very nice paper by Hairer, Lubich and Wanner on this topic, it is still on my 'have-to-really-read' list,;jsessionid=3CBD0B978747565DAE7A51ECA0E3FCC9?doi=

(chapter 5 is on energy conservation)

Physics Forums Values

We Value Quality
• Topics based on mainstream science
• Proper English grammar and spelling
We Value Civility
• Positive and compassionate attitudes
• Patience while debating
We Value Productivity
• Disciplined to remain on-topic
• Recognition of own weaknesses
• Solo and co-op problem solving