2001-08-08: Comments from Calandryll
Comments from Calandryll
Aug 8 2001 1:58PM
This was originally posted to Development at uo.com. [1]
Artificial Intelligence and Monsters
The dictionary defines artificial intelligence as: "The ability of a computer or other machine to perform those activities that are normally thought to require intelligence." In some industries, artificial intelligence could allow machines to perform complex and/or dangerous tasks, thus saving humans the risk or trouble of doing such tasks. But what does AI really mean when we are talking about massively multiplayer monsters?
First, I think it's important that we define what the purpose of AI is for massively multiplayer gaming and, more specifically, for the monsters in the games. Is the purpose of today's gaming AI to replicate human thought? Not really. The above definition is nice, but it doesn't really apply to what we do. Some might say that AI is needed to make monsters more challenging. To me, that's not taking it far enough. For instance, I could make a new monster that has 25,000 hit points, 102AR, 500.0 magic resist, is immune to poison, and does 99 points of damage with a ranged weapon that hits 100% of the time. That monster would certainly be challenging to kill, but there's no artificial intelligence in the monster beyond its ability to target and attack players. It's not a "fun" monster.
And that to me is what AI in these games is really about...making monsters fun. That may sound rather simplistic, but sometimes the simplest explanation is the most elegant. If a monster is not fun to fight, nobody is going to consider it a successful monster design, regardless of how challenging it may be. Fun does not always equal challenging and, more importantly, many players do not want monsters that are as intelligent as players...otherwise, why not just fight players? They want monsters that are fun. Sometimes a fun monster can be something that reacts to your actions, or perhaps it works together with other monsters of its kind to defeat its enemies. Whatever its abilities, if the monster isn't fun, then all of the artificial intelligence in the world is meaningless.
To illustrate the kind of AI I am talking about, I'll give you the example of the savage shaman. Each time a savage shaman is hit, she has a chance to initiate a tribal dance. If she initiates the dance, she searches the area for other savage shamans. If she finds three or more savage shamans within a certain radius, the shaman and all those near her begin to dance. When they finish the dance, the savages cast a powerful group spell that can cause harm to players or even heal all savages near the casters. Needless to say, letting them cast these spells could spell doom for would-be adventurers. But, if you are able to slay enough shamans before they finish their dance, then you can interrupt the spell before it is cast. Or, you can run away from the savages and if you escape the spells' range, avoid the damage of the spell. This kind of AI causes you to react differently than you normally would, without creating undue challenge, and gives you choices for how to react. It also has visual cues, letting you know something is about to happen and giving you time (although not much...you’ve got to think fast!) to decide what to do. Do you try to kill the shamans before they finish dancing? Or do you run away, hoping you can escape before they finish? Or perhaps you just continue fighting, preparing for the spell, knowing that your friends can handle its effects. If savage shamans were nothing more than white hued humans that cast spells they'd be no different, for all intents and purposes, than evil mages, orc mages, or any of the other spell casting monsters with similar stats. It's not their look that separates them from the other monsters in UO...it's what they do.
Of course, none of this takes into account the other reason players like to fight monsters...loot. No matter how cool a monster is, if the reward is poor, people will be less likely to interact with it and hence, the monster will not be considered a success. But how often can we keep adding the typical 100-250 gold and a 10% chance of some rare new item on every new monster? No, there has to be other kinds of rewards to engage players with new monsters. Perhaps fighting these monsters is necessary to achieve some part of a scenario, like the savages in the previous scenario. Or maybe, the monster isn't aggressive at all, but rather, its actions give you evidence (like the frightened orc) of changes in the world. New ideas of intangible rewards, like titles or even abilities unavailable through normal game-play, should be explored.
I don't believe that we, as designers, have even scratched the surface of monster design with massively multi-player games. Aside from the way they look, too many monsters are nothing more than different variations of hit points, damage, and armor values. The team for Ultima Online is going to be trying out some new things over the coming months to help change that. Hopefully we'll present you with some new, exciting monsters...and of course, some fun.
Jonathan "Calandryll" Hanna Designer, Ongoing Content
References:
