>then people who have faster computers will have an advantage.
How about requiring code submission, but only running in-house the top-competing(top 20? 100?) bots (if CPU parity is a defining aspect of the competition). My guess is the majority of submissions are just simpler student coding experiments that aren't designed with brute-force CPU in mind. If they are primarily faster CPU-propelled, then they are limited as such if they make it up the competition ladder to the in-house testing.