[Taylor]

I'd suggest Aim High (2006) as an example of a great game. This allowed for great diversity in robot design (no two robots really looked the same); attainable goals for every level (high goals & floor goals & ramp climbing); easy to understand (score balls in goals for points, high goal was worth more); alliance strategy (backbot selection was important); exciting endgame and worthwhile autonomous mode; human player interactions.
Improvements to be made to the game: More reliable real-time scoring, improved ball-return system.

(The argument could be made that the bridges in 2012 allowed for more coopetition, and I would agree with that. However, low-resource teams had more opportunities to be beneficial to their alliances in 2006 than in 2012.)