Programming and software development for engineers

Final project grading guidelines

• We'll run your agents as bots (i.e. equivalent to observer mode). Each run will consist of multiple rounds, running at a high tick_rate for the equivalent of 5 minutes.
• Run 1: We'll run your agent in a room by itself. You will receive up to 1 points, based on your performance. Run 2: We'll then run your agent 1 vs 1, 1 vs 2, and 1 vs 3 against a random sample of our staff's bots. You will receive up to 5 points, based on your performance.
• We'll review your code structure and comments. We want to see class-level and method-level comments. You will earn up to 1 point for how your code is written.
• We'll review your one pager, earning you up to 1.5 points. Your one pager can explain your current algorithm, ideas you tried, issues you faced, or any other insight you gained while implementing your agent. Your one pager can be in text or graphical form and must be submitted as a pdf, one A4 page.
• Your final grade will be the sum of the above points, capped at 6. (For context, our thinking is that beating the staff bots ~95% of the time should automatically earn you a 6, irrespective of your code and one pager. Beating the staff bots ~50% of the time with good code and a good one pager should also earn you a 6. Finally, if you rarely beat our bots but your code and one pager are good, we want to give you a 4). Note:
• We understand that testing an agent is difficult. We therefore don't expect you to write tests. You are welcome to test a subset of your methods, if you think it's useful.
• We will not be grading students vs students. I.e. your goal is to beat staff agents, not each other.
• You can test your code against the staff bots by connecting to the lab. We currently have 13 agents of varying difficulty and each connection assigns you a random set of staff agents.
• This is a two-person project. We expect each pair to submit a unique solution, i.e. there shouldn't be any code sharing between teams. Make sure you record your team.