You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When it comes to making an AI and determining if it is better, there is no better way than running it against your old AI. Obviously, this isn't foolproof, but if you beat your old AI, you must be doing something right. Being able to do this to you last version, would be nice, but that can make things weird if you are trying to fix a bug as the new last AI will have the bug. This is why I think it would be awesome if there was a system in place to upload an AI as your testing opponent. You can throw up an older version of your AI, or you can keep it as a shell AI, but the point is it would be a reliable way to test your AI quickly against a more constant foe, thus allowing you to measure improvement more accurately.
The text was updated successfully, but these errors were encountered:
Thinking through this idea, it might have just a bit too much overhead to implement fully on the Arena/Website's end. Since we don't really support playing yourself by default, this would be something easier accomplished via a bash script or something like that. It just may not be the best fit for the website, since it's mainly meant for trying your code out against other competitors.
When it comes to making an AI and determining if it is better, there is no better way than running it against your old AI. Obviously, this isn't foolproof, but if you beat your old AI, you must be doing something right. Being able to do this to you last version, would be nice, but that can make things weird if you are trying to fix a bug as the new last AI will have the bug. This is why I think it would be awesome if there was a system in place to upload an AI as your testing opponent. You can throw up an older version of your AI, or you can keep it as a shell AI, but the point is it would be a reliable way to test your AI quickly against a more constant foe, thus allowing you to measure improvement more accurately.
The text was updated successfully, but these errors were encountered: