Other:
A valid (rigorous) design process.
Inexperienced mentors, some half-hearted students, mediocre facilities, support, sponsors, scouting, etc can all be managed, but without a solid design, (produced bya good process) the robot will generally not perform well - and nothing can make up for that.
This was a tough question Ed, thanks for bringing it up.
I thought really carefully about this, because (as noted) all play a very important part in a team's success. But all of those factors can be fairly sub-optimum and you can still end up with a great robot.
Just below design is "Driver Practice", because a superb drive team can overcome a lot of bad robot. This sort-of implies a practice (or prototype) robot.
We have built 2 robots for the past few years, and it makes a huge difference (and thanks to Paul Kloberg for pointing that out to us). The first one is "good enough" and finished around week 4, then we take what we learned from building that and make our 'production bot' in the last 2 weeks. They don't have to be identical (pretty close though), but they always must use exactly the same software.
The advantage to continuing development on code and driver practice after ship date makes the difference.
now now, there is no such thing as reading the manual too much. I read it every night just before I go to bed. - 663.keith [more] 
Threaded Mode