It may seem unnecessary, but I always thought about this possibility to test the entire team or just one Pokémon. You choose the team of the Computer, so you can test your skills and weaknesses of your team, before testing, officially, on the server. I think that would raise the level of the competitions, because you could correct minor errors in your new (or old) team without showing the team for everyone and bringing the team to perfection. Sure, it may not be a priority now, but I think it would be very interesting. :rolleyes:
The intense amount of coding required to create a good AI is way too much. The AI would have to know how to use each of the sets you create for it. For instance, the ShbPetaya Empoleon set(which works extremley well on my mono water team btw) may be very easy for you to use, the computer may not know what to do with the givens. Even if a mildly competent AI is created, it could not get into the psychological game of switches and the like. I mean, developers, by all means correct me if I am wrong, but from little coding experience I have, AI is by far the hardest to code Other than the fact it takes lots of coding, it's a good idea
It would actually be quite hard because we would have to develop an artificial intelligence to deal with each and every move for every pokemon and that is no small feat. If we have a larger team or a skilled AI developer I could see this happening but it would take a long time.
You could always have a really dumb AI that only picks super effective moves or moves with the highest damage possibility and randomize it a bit. Not really "intelligence". I think it functions similarly in game with different levels of "intelligence". ie; younger trainers mostly go random, Gym Leaders/Elite Four/Champion and Ace Trainers would go for higher damage output with slight variations. It really couldn't predict much. If there were such a system, it'd be nifty whereas you could potentially have an online game with NPC opponents. Definitely not necessary, though.
Yeah, that seems like some heavy scripting and we probably don't want this to become skynet. I'm sorry, I couldn't resist.
Kioku, I agree we could do that, but what he wants is a way to test his team against a computer. This means we would need to provide a standard player, which atleast somewhat knows the metagame. In this regard, the AI would fail. On another note, the AI would be way to easy to predict. On a final side note, does anyone know what the pokemon games are coded in?
It wouldn't really be that hard doing an AI. I'm just afraid of the load on the server, as it would be as "heavy" as many battles together (that would depend of the level of the AI). Or, if you mean making a bot that has an AI and is separated from the server, then that would require intense coding.
If I understand correctly, the main problem is that the server controls the game logic (and rightly so!). Any AI would have to interact with it and query the server... alot. To answer the earlier question though, AI is easy to program. Smart AI is a bit more difficult... challenging AI is a whole project on its own.
There's not really much point in discussing this as it's been said that it's not a feasible suggestion. Closing for now.