Here are the five things I need to be a productive mathematician.
1. Time to think
2. Time to think
3. LATEX (I find most of my errors while typing my work)
4. Someone to talk to (it is great having a collaborator)
5. Coffee (I have no memory of not liking the taste of coffee)
[See Christoph Niemann’s coffee art]
David Richeson: Division by Zero 
I’d add #6. Sleep.
For myself, an evening spent flailing around on a conjecture is fairly likely to lead to a flash of “Ah-HA!” the next morning as I shower the cobwebs away. Somehow sleeping on a problem gets the creative juices going in a way that nothing else does.
But then again, I suppose sleep is just another variation on your #1 and #2.
Good point! It can go either way, though. A “solved” conjecture one night can come crashing down from an easy counterexample during a shower the next morning.
I’d be interested in your ideas as to how to procure #1 and #2, especially under teaching duties averaging 12-14 hours per semester.
Hmmm…. Yeah, I’m still working on that. I have a sabbatical coming up in a year and a half…