Cool stuff. I really like the jump and race arenas.
The maps and components will be based around the following game categories:
1: Jump
2: Race
3: Survival
4: Other -> Fight, Bridge/Climb/Obstacle Coop, Save-The-Robot, FLL/FTC/FRC, Sniping...
As far as game modes go, I obviously think you should follow an ingame setup similar to mine (
posted in the Arena Ideas thread). As it stands, KOTH is the only game mode that doesn't award points for attacking other bots so it is ideal for task-oriented game types (like racing, jousting, and soccer).
I have come up with a few different methods, though I have yet to act out and decide on one. I have only run some tests with them. A simple way, for a map like a drag race, I could just have the game add points ~ every 200 milliseconds to the robots, and then immobilize them at the end of the track. Once immobilized they would not be earning points, therefore their time/score would be locked and can be compared to the other robot once it has finished. In the end the two players will compare their scores to see who has the lowest score. It should be pretty easy.
This works but there is no way of checking the bot scores unless you keep track of it on your own. You'll also run into problems with the immobilization counter if you plan on disabling the bots rather than destroying or eliminating them (obviously, you could disable the timers, drop the bot, or use a force/push but these solutions aren't ideal). I can't recall if Clickbeetle ever said what route he was going to go with his race arenas but here are few things that I've done:
1. Eliminate or destroy the bot when it reaches the finish line and use a hack to declare it the winner after the other bots finish (easier than it sounds).
2. Eliminate or destroy the bot and award it negative points equal to the total time it took to finish (I can't remember if you can have a negative total but if not, just subtract from a large number).
3. Eliminate or destroy the other bots.
Have you thought about how you are going to handle the AI? Once again, I've managed to do it in three ways:
1. Custom AI.py - if done right, the AI would be compatible with all the game modes (simply by checking to see what arena or game type was started).
2. Modifying waypoints from the Arena.py - you can feed the AI new target locations by feeding it modified waypoints (preferably using fuzzy locations, target planes, or a navigation mesh).
3. Injecting code from the Arena.py - you can replace methods within the AI.py at startup to cause all AI.py's to be compatible with the arena (combines nearly all of the benefits of solution 2 with some of the benefits of solution 1).
Edit: 4. Registering target zones - Similar to how in a King of the Hill match the AI goes to the top of the hill, you can register/unregister a zone for each location that you want ALL the AI to navigate to. This doesn't work well with multiple AI and I can't remember offhand if it is possible to unregister a zone (I just remember Mad or Click saying something about this).
Some other ideas: PortalGun/teleportation in a map, map with gravity reversed past a certain z value, capture the flag...
Teleportation is really messy in RA2 (and not recommended). The problem is that there is no builtin command for it so you would need to simulate using the limitted number of commands that RA2 does provide. Here are a few different solutions with drawbacks:
1. You could toss the bot to the other location using the force, push, or gravity commands but the problem is stopping them and allowing for a graceful landing (even with a very fast tick interval, I haven't had any luck exerting an equal and opposite force because none of the commands are linear).
2. You could move the platform that the bot is on using a hinge or prismatic but you wouldn't be able to go very fast. This is especially true with prismatics because a glitch in how they were coded causes bots to be launched when they stop moving. Out of all the solutions, this has the best chances of keeping a bot upright.
3. Similar to solution 2, you could enclose the bot inside of an invisible shell and move the object into position using a hinge or prismatic. The shell would compensate for some loss of control but it greatly complicates things and still wouldn't prevent the bot from being inverted.
Show Posts
