[Al Skierkiewicz]

In no particular order...
1. Design your robot to play the game.
2. Design your robot to be robust and repairable.
3. Design your robot to complement teams that you will play with.
4. Design your robot to effectively carry out your strategy for play.
5. Practice, practice, practice!
6. Know what your opponents and allies are capable of.
7. Have extremely good luck.
8. Learn from other's mistakes.
9. Be ready to modify any of the above at any time.

We have learned more from losing and that drives our design procedure and how we interact with the robot following each match. We have lost when a driver has missed an opportunity, or zigged when they should have zagged. We have lost because we failed to know what our opponent was capable of doing or because we believed an ally could perform a task they had never accomplished. We have lost when a component failed or an electrical connection has come apart or when an assembly didn't quite achieve what it needed to achieve. We have lost to teams with better luck (for lack of evidence of anything already mentioned).
Don't be fooled. Teams have won with no CAD abilities, little shop tools, few mentors or engineers, little money, few resources, few students, little software. They have done the best with what they had and became very good at driving and manipulating the robot.
And when luck fails them, other teams will recognize their potential and pick them. (That is a big thank you to all the teams that have done that.)