5 ms·
The problem I tend to have with user created maps in RTS games is they often don't work well with computer AI at all. The AI in these games often relies on cer
by jiggy2011 14y ago
The problem I tend to have with user created maps in RTS games is they often don't work well with computer AI at all.
The AI in these games often relies on certain "trigger points "on the map for example "try and hold this area", "try and attack this way" , "build a base here" etc rather than having a truly generic AI.
When using user created maps these points are often either not put in or are not used in a way that is not as tested and thought out as those that ship with the game.
This results in AI that either sits there and does nothing or will just play very badly on a particular map.
Of course most advanced players are not interested in playing against AI, but it is a good way to learn a new game or to try a particular technique etc.
- lmm 14y agoReal AI is hard. It's not that developers don't try, it's that they can't make it better. About the best the developers can do is give players access to the same interface they have. If you look at the StarCraft community you see amazingly complex AI scripting using what's fundamentally a very simplistic trigger interface. But it's only the top epsilon% most polished maps where the creators put in all the effort needed to make that work.
- b0rsuk 14y agoMagarena and Unity of Command may not have a true AI, but it's still very good. Have fun checking them out. Unity of Command is an old-school looking hex-based WW2 game. Magarena is a free program to play Magic: The Gathering. It has a few downsides and doesn't deal with fuzzy logic, but most of the time it plays like a bastard. If you underestimate it or hope it doesn't notice something, you'll regret it.
- jiggy2011 14y agoAI always seems much better in turn based games than in real time games. I guess because turn based games are more able to say "stop everything until I recalculate all of the path finding" whereas a real time game has to deliver so many game updates per second and will maybe have to run AI in a separate thread which might not yield a good solution until the AI has already had to make a decision based on stale information. Chess AIs are often scary good for example and FPS AIs are notoriously dumb.
- b0rsuk 14y agoWhen you bring in FPS games, keep in mind single player in most of them don't treat AI players as first class citizens. This is what C-Evo developer calls "AI liberation". Examples of FPS games with decent AI are Unreal and Unreal Tournament. The latter also has the best bots to date, I think.
- jiggy2011 14y agoTrue, UT had reasonable AI but it did always seem to perform much better on deathmatch where it could just run around shooting things than on CTF maps where you could quite easily camp outside of your base and just snipe it when it tried to come in. Of course AI is getting better now that we can offload stuff onto other cores without tying up the game loop so much.
- thaumasiotes 14y agoI don't know to what degree the UT bots qualify as AI players... their strategies are coded into the level, not the bots. (disclaimer: I read guides to making UT levels, but never actually made one. Details may be misremembered. However, the fact that two levels with identical topography might be treated entirely differently by the same bots (owing to different bot-path placement in each) doesn't leave me filled with awe at the bots' decision-making.)
- neumann_alfred 14y agoPlacing paths manually or having them auto-generated by the map editor is standard practice. It's mostly because it would be expensive/silly to calculate at runtime -- calculate it when it changes, save it in a format bots can parse; the bot AI still makes decisions based on that path. How is that different from first calculating it?
- lmm 14y agoCan the paths be usefully autogenerated (at whatever computational cost)? Or do they have to be done manually? The latter puts us a good deal further away from "real" AI.