Tide's Horizon

Intelligence in MMO's (or Why I Can't Have My "House MD" MMO)

Among the many things that must drive MMO designers nuts is the repetition of already well known, or well worn ideas and problems that players some times justcomeupwith. But all of which designers are supposed to be fixing. Or expected to fix in next-gen-MMO Teh Quickening! or whatever. Problems like permadeath. Another gem is mob AI.

On Quarter to Three there's a discussion about why AI sucks in MMO's. A lot of it is wondering about why most mobs don't run away, don't beg for their lives, don't negotiate, don't come back in the next gaming session with their cousins and their cousins' friends to whup yer ass, or really do anything ever "intelligent". Predictably, most followup's have been anecdotes about how WoW is better, or how things were really better way back when in UO or EQ. All sarcasm aside, there is a real problem here. But dumb mob AI is just the symptom and not the cause. In games we don't expect, for instance, the shoe in Monopoly to act "intelligently". So why should MMO's be any different?



Lum makes an entrance and points out the following:

Gamers say they want better AI in MMOs. Data mined statistics show that the most popular monsters in any MMO are the ones with the most brain dead AI, preferably bugged into virtual unconsciousness. CPU load is another issue, but not as much. Many MMOs have very intricate AI encounters - but players tend to hate them because good AI adds a concurrent risk. While saying "hey, it'd be cool if guards responded when their buddies are attacked" while discussing the subject, it's entirely different when you attack a monster and are suddenly fighting off 10.

It's a classic game design challenge - make the player *feel* as though they have completed a difficult challenge, while keeping it easy enough for them to actually accomplish.

This is important -- MMO gamers want more of an interesting challenge, but they only stick to the uninteresting and predictable challenges. Why the contradiction?

A lot of what I end up saying about MMO's always comes back to what kind of advancement model there is, and mob AI isn't a problem that's any different, unfortunately. What Lum is credibly pointing out indirectly is that MMO players always approach tasks and encounters with an ROI model in their minds; namely, "Is this encounter worth my time?" and that's because every single mob in an MMO is (to quote Lum) "a mobile bag of improvement". Since every popular MMO has an advancement model that's based on consumption ("killing") there is a need to have "mobile bags of improvement" instead of unpredictable or even dramatic opponents. The reason is because the game design has a progression metric that is predicated on players consuming content and being graded on the amount they consume. Levels are based on the quantity of dumb-AI driven models that are "killed". This is not going to change while consumers expect it.

You can't have "intelligent" AI in an MMO until you have an MMO that requires intelligent play. Repetitive, "devotional" play styles don't need deep creativity or real analysis, because they are reactions to designs that are just timed end-game pathing puzzles. That is, "How fast can I get to the cap to play the end game and access all the content in the game?" The time to the final return on my investment (time, subscription and license fees, any spoilers I've written, etc.) is what counts, and any constraints that delay that or diminish it or even risk my ROI goals won't be accepted. That includes mobs that act in an unpredictable way.

It's fine to propose interesting or somewhat novel AI whereby the models "act" differently (and this was brought up in the thread with SWG in mind), but they're only ever going to be different kinds of motion capture. The AI can't change in a regular MMO that's consumption driven. Whatever twists are devised for how to do aggro or chaining (social mobs) they are still all for the same vector of delivering content to the player quickly and efficiently for them to consume.

FWIW, the ultimate thing for me with MMO's is to wonder about how to change advancement. A good debate about this occurred last year over on Raph's blog whereby he went through "why levels suck". But a great rejoinder by HRose to a post Raph made about a medical MMO illustrates the problem really well.

Raph proposed a game where you had to "heal" patients; HRose argued it was the same old, same old. There was a difference in interpretation, but HRose was correct in that whether it was healing or killing dumb AI models, the game design seemed the same of consuming your way through the content. In Raph's defense, he was interested in the style of the healing and the "how" of the consuming, while HRose was (like me) concerned about the "why" of the need to level up through how many patients you can heal.

Why this is interesting is that you can't have a "House MD MMO" until the problem of how to accrue advancement points in a way that doesn't depend on consuming "mobile bags of improvement" is resolved. In this example, House is a detective story driven program where medicine is the "how" of the show, but diagnosis and investigation is the "why". (The main character even lives at a "221B" address.) His staff always worries about the genetic background and habits of their clients (patients), often going so far as to break into their homes to look for evidence and clues to the patient's illness. How can you build an MMO on that very popular formula when your advancement model is consumption? The main character hates dealing with patients, and they only have one "client" per show... how can you level-up?

This problem isn't going to be solved soon, and it won't be innovated probably by the big MMO providers. Or even the webby virtual world services. But there is a compelling need to do so, because so many of the great franchises people love (e.g. Sherlock Holmes, Dr.Who, Alice in Wonderland, Wizard of Oz, etc.) don't depend on consumption. They aren't all pie-eating contests, and certainly there’s no killing in them. You can't make a Harry Potter or a Lost or a CSI or a House MMO until someone makes a progression metric and method that's not based on consumption. Each of those franchises are about investigation and puzzle solving.

Advancement by intelligence is way more sticky than advancement by repetition. It's also way more personal, and so, compelling. And because the game can appeal personally it also can be commercialized more deeply. As I keep saying repeatedly, the people who solve this problem will probably own the next generation of MMO's.



[Mar 16 Update:]
There were several worthwhile posts around these core mechanics by Ubiq (Damion Schubert) following his rountables at this year's GDC. See below. I should've also referenced his Sept.'06 AGC presentation where he discusses the weird problem of people wanting new MMO's, but having still a general reluctance to also try anything new.

• MMO Roundtable Day 1: Why Fantasy?
  
• MMO Roundtable Day 2: Why Combat?
  
• MMO Roundtable Day 3: Why Classes, Levels, and the Grind?

I'm not sure exactly in all this where Ubiq's mind is, because on the one hand he's a great supporter of innovation, but on the other he's now had at least 3 presentations where he defends the combat/fantasy/DIKU model of MMO. I think at the end of the day he's really trying to emphasize craftsmanship over half-ass title launches. And kudos to him for that. But I still wonder how much the combat/fantasy/DIKU model and trope is popular because it's "popular", and/or because it's easier to produce and manage. I suspect it's about 50/50.

Labels: advancement, Game Design, MMO, MMORPG

posted by Tide @ Thursday, March 15, 2007