Evaluating designs is a critical part of being a good engineer. After initial brainstorming/concepting the pros and cons of each alternative must somehow be quantified. For our cars it's generally a compromise of speed vs feasibility (cost, time to construct, etc).
Quantifying the benefit for each is the difficult part, showing how critical your skills in the design-of-experiments are. Many times this step in the design process is left out and the coolest option is chosen (or the benefits of each are swayed by the designer towards the option they like most). This is dangerous because design judges often see through the "cool factor."
Proper decisions must be made considering test data (stopwatch, data acq, flowbench, etc), theoretical analysis (CAD,CAE, CFD etc), manufacturing practices, and material costs. Only when all feasible options have been explored, can a proper engineering decision be made. If you can prove that you went through this process then no judge can tell you that you were wrong (unless you didn't do your analysis correctly, for example). Ideally this is where you prototype and re-evaluate your decision. Depending on the time and cost this may not be feasible (we're building prototype cars, right?)
After a design is chosen and it's been built, testing and validation is a NECESSARY step in the design cycle. No design is complete without determining if it is better than previous designs. Here is where judges will kill you. You can show why you made the decision initially, but if you can't back it up and explain why it's on the car, then it's still no good. From experience, I can vouch that most judges would rather see that you tested your part and found that it doesn't work as well as a previous design, than just slapping it on the car and calling it a day. It's even better if you have the time to make modifications and improve the design.
Finally, don't forget to return to the initial design criteria to see if you meet it. Any design presentation that shows that these steps were taken is top notch in my books. Pat, what do you think?
It's all part of the design cycle...
test, design, analysis, prototype, test, design...
