I've noticed that the agent seems to persist throughout the tournament, since it is only created once at the start. This means that it could remember things between games in the tournament.

When the projects are marked, are they tested using the same tournament code that we have or some other way, and if so, are we allowed to use this ability to remember things between games? (I suspect this is what the game_outcome method is for, since otherwise it can't see what it would be used for, but I could be misunderstanding it.)

Thanks

ANONYMOUS wrote:

I've noticed that the agent seems to persist throughout the tournament, since it is only created once at the start. This means that it could remember things between games in the tournament.

When the projects are marked, are they tested using the same tournament code that we have or some other way, and if so, are we allowed to use this ability to remember things between games? (I suspect this is what the game_outcome method is for, since otherwise it can't see what it would be used for, but I could be misunderstanding it.)

Thanks

No, you should not assume that your agent will survive between games. The testing code provided is not exactly the same as how your code will be assessed, and game_outcome is provided so your agents as you are developing them can know the outcome (useful for learning techniques, etc.). Your agent as submitted should be ready to be dropped into a game with no "warm up time", and may be assessed as such.

Besides, the only thing I think you could do with this information between rounds is metagaming by trying to learn how other agents you are up against play. This may work at the testing scale, where there are only a handful of agents you are competing against, but in more thorough testing against more effective agents, it can't help you.

So I guess you are welcome to implement this, but I do not expect it will actually make your agent more effective in any measurable way, so probably not a priority!