Tuesday, June 30, 2009

Emergent Solutions vs. Exploitative Solutions

In my last post about minimalist design, I didn’t go very deeply into the implications of minimalism in MMO design. I’m going to work on the idea of minimalism and emergent behavior over the course of several posts exploring what emergent behavior is, what it means for MMORPGs, some suggestions how to make it happen, and the new problems it introduces. This post is about what emergent behavior is, how it manifests itself, and why theme-park MMORPGs don’t make it happen.

This article on emergent solutions in interactive fiction fascinated me. This is exactly the kind of things I’d like to see in MMOs. Instead of having the goals be resting points set by the developer, though, I want players to set goals for one another and interact to find solutions to the problems that stop them from reaching these goals. This goes much deeper than “I want to be more powerful. I guess I’ll kill a thousand rats. I’ll do that by swinging my sword at these rats until there are corpses on the ground and experience points on my character sheet.”

An emergent solution is a solution that the developer did not premeditate as being a possible solution to a problem. You allow emergent solutions by providing a rich game world with sets of interrelated operations that players can perform on different items in the world, modifying their properties and allowing an explosion of possible paths of action. The author of the article I linked to earlier mentions a great example of an emergent solution: to get the seeds out of a seed bag, the player killed a rat, cut off its tail, froze that tail in liquid nitrogen knowing that it would become brittle, shattered the tail into sharp shards, then cut open the bag with one of the shards.

Exploitation as an Emergent Behavior

The primary kind of emergent solution behavior in the current generation of MMOs is actually exploitation. The game puts the players on rails towards completing certain tasks that are supposed to be at a certain challenge level; to compromise the tasks by reducing the challenge level through exploiting the flaws in the game logic surrounding them is exploitation and poor behavior. The designers have thought up the problem and the solution and you had better solve that problem with that solution! It doesn’t matter if there are other feasible solutions, because the game isn’t about actually solving problems, it’s about receiving the rewards—or at least that’s what the developer is telling you through his design.

But exploitation is emergent behavior. If you see game systems only as arbitrary sets of rules that stand in the way of accomplishing what you desire, then there is no difference between using a wall hack in Darkfall to farm mobs imperviously and cutting off a rat’s tail, dipping it in liquid nitrogen, shattering it and using the sharp shards to puncture seed bags so your character can feed itself. Game rules do have meaning, though, because they relate to what we do and what we’ve seen done. People can’t walk into walls to avoid being hit by missiles, so this is considered an exploit. A person could conceivably use the shards of a frozen rat-tail to cut open a seed bag, so it’s considered an emergent solution.

The Leakiest Metaphor

It’s important to understand game rules as metaphors for real life causal relationships. In MMORPGs, this relationship can become obscured by the gulf that currently exists between what a player should be able to do if the metaphors hold and what the game allows the player to do. MMORPGs have extremely limited player-world interaction schemes. The player has a tiny vocabulary of actions he can perform and few of them have any lasting effect on the game world. The metaphors only apply at a very abstract level: you can fight, make stuff, and get raw materials out of the earth. Those three actions vaguely mirror their real-life counterparts if you squint very, very hard.

The good kind of emergent behavior occurs when you seal up those places where the metaphors squirm and fail. You don’t have to directly model every single part of a real-life process to seal these gaps. You just have to do a good enough job of designing and implementing with minimal-impact bugs a metaphor that is uniform in its depth and actually engaging to play through. If the metaphor is inconsistent or too shallow, players will rub up against those things that it seems they should be able to do but they can’t because the game arbitrarily seems to frustrate their attempts.

7 comments:

Psychochild said...

I talked a bit about cheating at a recent conference. The line between "cheating" and "clever" is indistinct.

The main issue is keeping the game fun for everyone. It may seem like the developers are being arbitrary (and they might be), but most of the time a strategy that is too effective is removed to preserve the fun of the game. This could be to avoid cheapening the experience (especially of those who already did the content), or maybe avoiding "you must have a party composition of X" situations that just aren't fun for people who fall outside of X.

That said, there are probably times when developer ego at creating a challenging encounter gets the best of them. It really kind of sucks for a player to have an ability nerfed because it was used a specific way. So, developers should really take a hard look and see if a situation is truly an exploit or a bit of cleverness.

motstandet said...

MMOs are filled with these emergent solutions. No, players aren't freezing the tails of rats off, but they are still using the rules and items in the game in ways originally unintended.

A major source of gil in FFXI used to be to fish up rusty items (usually a sign of a bad fishing location). Instead of destroying the Rusty Cap or selling it for next to nothing to a vendor, players learned that they could get their Blacksmithing up to the mid 50s and use a Light Crystal (the Restoration Synthesis) to turn that cap into a more valuable item which vendors bought for a pretty penny. [This was eventually removed since it was very easy to bot.]

Or how about in WoW, before Warlocks were able to Summon into instances: you activate a Summon on a player already in the instance, that player could Hearth to fetch items or repair, and then accept the Summon to be transported back into the instance.

What you are asking for isn't that revolutionary. To me competitive games have always had emergent solutions. What else would strategies and counter-strategies be? TF2 Engineers use their buildings to get to previously unreachable places; players use items as projectiles in Super Smash Brothers rather then using the item's ability; DotA players use the fog of war and tree lines to "juke" a pursuing enemy; WoW players theorycraft for hours determining the optimal skill rotation and talent builds from among hundreds of possible combinations.

motstandet said...

This post has some more examples of unintended play.
http://blogatelle.wordpress.com/2009/06/20/unintended-play-why-does-it-threaten-people/

Psychochild said...

Most of the examples you gave are game hyper-optimization. The FFXI rusty items is probably something that was intended to be in the game, but the developers probably made the newbie mistake of thinking "nobody will seriously do that!" Well, people will, and it's easy to bot to boot!

The TF2 example is a bit different since, in theory, any player could choose that class and do that "exploit". Same as in DotA. As I commented on a more recent post, an "equal" playing field like TF2 isn't something people will pay a subscription for, though.

But, not all "emergent behavior" is good. An early WoW exploit involved rogues stealthing to the end of a raid, pulling the boss, then getting summoned back to the beginning by Warlocks. The boss then ran through the whole instance to catch up with the rogues and the group just killed the boss at the entrance without having to clear trash. Pretty clever, but this causes problems beyond developer ego; if this strategy had been allowed to stay, it would mean that most encounters would have to have a rogue and warlock to be played optimally. This probably would have shut mages out of many encounters, because why bring a mage when a warlock can do the exploit? (Actually, that was the attitude for a while even without the exploit being allowed, so it would have been worse.)

In this situation, if I had been in charge of WoW and found that exploit, I would have thought it was clever but would have fixed it anyway for the reason I state. Yes, it's clever, but it causes problems for a lot of players in the game. There are worse fates than players having to play through the game as a competent designer intended in order too make more people in the game happy.

motstandet said...

Wouldn't most emergent solutions surface in cases of optimization? Unless the player wanted to try his hand at a Rube Goldberg machine.

Those class-based, compartmentalized progression schemes are difficult to work with in terms of balancing, especially when you ask the player to invest weeks and months into that progression. If all the players had a way to trigger that exploit, would it still be an exploit?

Tesh said...

Put another way: "If everyone is doing it, is it still wrong"? It depends. If it's against the designer's intent, I'd have to say yes, it's still an exploit.

In that case, though, there's probably just something wrong with the implementation not lining up with the intent, so it's something the dev needs to fix, and be grateful that the players pointed it out.

dancingelephants said...

I like where this series is going and the rat analogy is a good one. The concept of the exploit comes about because the designer strictly wants to control everything about the game; as in chess. Layers and layers of rules get added; mostly as special cases to prevent "exploits"; which are just novel uses of the rules. Consider that technological innovation is essentially "finding an exploit" in the ruleset that is real world physics.

Perhaps Bartle's offhand remark that worlds should implement just enough naive physics to make the world believable has merit. Then you only need a very lightwieght ruleset. http://dancingelephants.wordpress.com/2009/02/05/lightwieght-and-heavywieght-rulesets/