For Storm II, the reverse of Mario.....

- Define what you want your robot to do at the highest level possible (we started with the word 'win')
- Break that initial statement down into what you think is required to achieve that goal (for us words like maintainability, reliability, ease of access were all things we considered in being required to win, long before we thought about weapons - they are a secondary thing)
- Understand the hardware that you want to build your robot out of, research, research, research. Attend some live events (http://www.robotslive.co.uk and http://www.roamingrobots.co.uk or http://www.robochallenge.co.uk) before you buy anything as having attended you might find you want to go back to stage one again
- Then... and only then start working out what needs to go where, and let that drive the overall size of the robot. Just propping parts up against each other as you aquire them is a far way to go, and never underestimate CAD (Cardboard Aided Design).... or if you fancy teaching yourself, Computer Aided Design

Only after all that, will you start to have a feel for how big something will be. But my honest advice is don't start with a pre-concieved idea about what the robot will look like, work out what you want it to do.