April 15, 2004

A New Game Model: Bots, Nurturance and Solving the Grind

It’s kind of ironic that EverQuest has just celebrated its 5th birthday, still going strong, and one might have imagined 5 years ago that we would have much better MMORPGs by now, but we don’t. Almost every game that has come out since EQ has been an EQ-clone. Actually, what’s more ironic is that EQ2’s biggest competitor will be EQ itself. Notable exceptions (A Tale in the Desert, Sims Online) have struggled to gain a sizeable player base. In a way, everyone is waiting for something innovative to come along, but it’s not clear whether anyone knows what that is.

Not that I think I know what the next big thing is, but the following is an idea that came to me after several weeks of playing Ragnarok Online. I’d like to briefly describe several interesting features of RO before moving on to describing my idea (impatient readers can skip ahead).

Notes on Ragnarok Online

RO caught my attention because it uses 2D anime-styled characters instead of 3D pseudo-realistic characters. It overlays the character sprites on top of the 3D environment pretty convincingly. One benefit of this is that you don’t walk into the city and lag at 1 fps (like in SWG). There are several unique game mechanic features in RO:


Transformations – There is something seductive about identity transformation on a very primal level. We are fascinated by metaphors of cocoons and butterflies, gray cygnets that become white swans, beasts that turn into princes. Of all the MMORPGs I have played (EQ, DAOC, SWG, RO), RO alone captures the seduction of transformation. Change is gradual in games like EQ and SWG, whereas in RO, you literally change appearances, abilities, and suddenly gain about 30-50% more hit and spell points twice in the course of the character’s career (in the addition to the gradual improvements). The visual transformation paired with the sudden access to several lines of new skills is powerfully motivating.

Repeated Cycles of Rewards – Instead of having a linear level treadmill, RO does something very clever. You have a base level and a job level which advance independently. When you change jobs (from Mage to Wizard for example), your new job level goes back to 1 and suddenly you can make job levels very quickly again (allows acquiring new skill points). The rewards cycles in these games are based on well-known behavioral conditioning principles. It entices you with instantaneous gratifications up front, and slowly eases you onto a treadmill that takes longer and longer before you get a food pellet … I mean level. So in RO, just when leveling gets slow, you get to experience the instantaneous gratification all over again. Just when you might have gotten bored of the treadmill, it becomes rewarding again.

There are two interesting sociological features of the game which I’m not sure are good or bad:


“Race”-less – There is only one race, which for all intense purposes might be called “anime”.

Gender Enforced – Your characters can only be the gender you put down when you registered for the account. So you can’t gender-bend.


====


Botting

Now here’s the feature of the game I found most fascinating. RO has incredibly poor network security, which makes it very easy for players to download and use bots (AI scripts that pretend to be the game client and literally do the work for the player). Botting is actually a well-known problem in the game, with several maps known to be infested with bots. Bot scripts are fairly sophisticated. Users can control:


- which map the bot should travel to and stay on (given any starting location in the world)
- what monsters on the map should be attacked
- what loot should be picked up (to save encumbrance)
- switching to different weapons when fighting different monsters
- resting when HP/SP is low
- using certain spells, abilities or items given a complex set of conditions
- teleporting to a random spot when a crisis occurs (ie. mobbed)
- heading back to town to store loot when encumbrance max is reached
- avoid kill-stealing or loot-stealing from other players
- auto-responding to chat and PM’s based on key words (ie. They can pass the Turing Test sometimes)
- disconnecting if a GM is spotted

Botting is perceived as a bad thing by Gravity (who owns the game) and by most players (because bad bots will kill-steal you), but in a weird way, bots fulfill a critical function in RO’s twisted rare-item economy. Many highly-demanded items (ores, cards, slotted items) have very low drop rates to the point where if bots were not around to harvest them, many of these items would be several times more costly and difficult to buy.

====


The Big Picture

Here’s the aspect of botting that intrigues me. Sophisticated botters could script a tank bot (knight/assassin) for leveling a new mage, or script a buff bot (priest) for leveling a new melee character. One could imagine the AI becoming smart enough such that you could script both and let both run at the same time. Or what if you could script your own dungeon-crawling group? A collaborative botting model introduces several new themes:


Character –vs- Entourage: The focus on individuality can be seen in video games. Most games have you controlling one character at a time, or as an overseeing god-like power without avatars (except in some RPGs). What if you controlled interacting characters instead of single characters in MMORPGs?

Persistent Characters: Interestingly, the only thing not persistent about persistent worlds are the characters that inhabit it. Your character is gone when you log off. So much for persistence. Bots allow characters to be persistent, in fact making the world more persistent.

The Nurturance Model: There’s something appealing about watching and being a part of a character’s growth – watching them grow from fledglings into masters of their disciplines, but that aspect is actually independent of the grind itself to a large degree. People get attached to plants and animals even though they grow and go about their existence when you’re not around them. Current game models try to get you bound to the grind (because the grind is what levels you up and grows your character), but that model severely cripples casual players. A nurturance model hooks the player on the character itself rather than the grind process.


====


The Details

But let’s take one step back and imagine what the game would actually look like.


It Takes a Village: Let’s say players start with a town with one or two characters at their control. These novice characters have the potential to level into a set of higher professions following RO’s model. The only difference would be that there would be more crafting professions that were more sensible (don’t require fighting to become a blacksmith or an alchemist). The goal of the game is to grow a large collaborative network of characters that complement each other and help the town grow.

Characters: Like in SWG, different characters complement each other well and need to work together to achieve larger goals. The difference here is that the player has to choose how to optimize his town according to his liking. Maybe he wants a more militaristic town that can take over other towns, or maybe he wants to be the merchant hub, or maybe the magic research hub.

Advancement: Apart from normal character leveling, we have two other ways of granting advancement to players. We can give them more novice characters to grow and we can give them more advanced botting scripts.

Control of Characters: Players can control the characters directly, but as they advance in the game, they gain AI scripts that allow them to automate their characters. Perhaps we would also build in several different areas of scripts such that certain players might choose to build great fighter bots while others build great harvesting/crafting bots. But in either case, players can choose to gain control of a character they want to play with.

Collaborative Botting: The complementary nature of character skills forces players to script collaborative bots that work together by themselves. In fact, this is where the real game is – how to optimize and get bots to do what you want by themselves. Players would create tank/healer pairs for hunting groups, harvester/bodyguard pairs for harvesting missions, and an elaborate production chain for mass-producing certain goods (for example, clay => bricks => wall enforcement, or ogre blood => red gems => spell research).

Focus on Strategy Not Grind: The semi-automation allows players to truly focus on the overall strategy of how they want their towns to develop as their village grows to about 20-30 bottable characters.

The Multiplayer Part: Different players collaborate or compete at a larger level now – two villages fighting over the same set of limited resources, or a fighter town that feeds resources into the manufacturing town and getting weapons in return. Diplomacy between multiple players becomes more complex and interesting.

A Battle Scene: Instead of controlling one character, imagine a player who is working with two other players to raid an orc city. They each have about 10 characters. One town will supply close-ranged fighters, while another will supply healers and mages, and the third will supply archers and fast cavalry. Many characters could be scripted to do basic tasks – like healing specific fighters, giving them priority over archers. We could also imagine a semi-automated group. The fighter town might have 8 bots that attack what 2 key characters attack – controlled by the player.

While You Sleep: In fact, the game makes it such that your village is working day and night even as you sleep or are at work. Your town is a living, breathing entity that persists in the world – harvesting raw resources, manufacturing enforcements and weaponry, battling hostile invaders, all according to your plan. You could set alarms to alert you to anomalies if certain strange conditions appear – healing potions running out very quickly.

====

Potential problems and answers

Run-away inflation: With a botted economy, there is the danger of out-of-control inflation. This can be controlled by having a closed player economy where players are not selling large amounts of unwanted goods to NPC merchants. Because everyone has access to automated harvesting, there isn’t a problem like in RO where some players have a lot more than others unfairly.

What about Socializing: This appears to be a very logistical and calculating game that at first glance seems devoid of socialization elements. But there is a lot going on at the inter-village level between players. We could also imagine many variations of non-combat oriented goals – a fashion tailoring village that produces unique clothing, an alchemy town that researches and sells high level spells, a ranger village where players can buy tamed horses for cavalry or exotic animals as pets, or towns where festivals are held.

Asynchronous Communication: At first glance, it may seem problematic that so many of the other people you would need to talk to have a good chance of not being logged on at the same time as you. This problem is solved by asynchronous communications like an in-game email system or a real-world IM system.

Bots are Bad: Because of the way bots are currently used, players think they can only do bad things like kill-steal you. But in fact, benevolent botters in RO could choose to set their bot to run around and heal or bless the players on a certain map. Indeed, we could build in altruism as a way to level up certain skills – like for priests who need to make a pilgrimage to another village and carry out priestly duties. The metrics and advancement requirements of a game are what guide players to do what they do. Like “A Tale in the Desert” or “There.com”, if you get cool skills from helping out newbies, then players will help out newbies. Altruism can in fact be engineered by the game mechanics.

So the question is: What’s your gut-feeling about this game? Would you want to play it? Would you add or change anything about it?

Posted by nyee at 8:21 PM | Comments (22) | TrackBack

March 26, 2003

Engineering Relationships

Introduction

The effects of game mechanics can be explored on many different levels. On the lower tiers, we can look at how the rewards system enhances or diminishes the appeal of the game. On the higher tiers, we can look at how game mechanics influence community-wide behaviors or phenomena. For example, it is probably fairly obvious that the game mechanics of an MMORPG affect the economy that develops within the game. If there are limited ways for the currency to leave the player market (through NPC vendors, death penalty, etc.), then inflation will eventually overtake the economy and be difficult to control. But it may be less obvious how the game mechanics of an MMORPG affect how relationships form and develop within the game. By comparing the game mechanics of EverQuest (EQ) and Dark Age of Camelot (DAOC), this essay explores how these game mechanics can shape the relationships that form in MMORPGs. While more theoretical than empirical, the ideas presented are all testable hypotheses. An understanding of the effect of game mechanics on social phenomena has an impact on the design of future virtual environments, as well as helping us understand how social context affects us in the real world.

Encouraging meaningful relationships is much more than just enhancing the communication interface. While clearly a necessary part of building relationships, having a communication channel doesn’t do any good if players aren’t encouraged to interact with each other. It also doesn’t do any good if players only interact for superficial reasons. To foster strong relationships, a game needs to provide players a large potential to interact and increase the likelihood that each interaction creates a relationship between the players involved.

Downtime During Fights

Forcing players to group to fight a tough mob is a typical way to get players to interact, and most MMORPGs make it very difficult to solo as the player’s level increases. But perhaps the amount of downtime between fights is also a crucial element in player interaction. DAOC streamlines combat and minimizes downtime during grouped combat. Mythic does this by making most buffs consume no mana, and by having fast HP and mana regen among other design elements. Typical grouped combats in EQ, on the other hand, are separated by pronounced intervals of downtime. Among other design elements, HP and mana regen are slow, and buffing a group consumes most of a cleric’s or druid’s mana, after which the group has to wait until the cleric or druid has regained that mana. Also, typical battles with a mob are shorter in DAOC when compared with EQ, and the rate of mob encounters is higher in DAOC than in EQ. In effect, what typically happens in EQ is that a group fights for 5-10 minutes and then has to rest for 3-5 minutes, while in DAOC, a group can fight continuously for long periods with relatively short rest periods. Even though players are together in a group and might be inclined to talk to each other, they can’t really develop meaningful relationships easily if there’s not enough time to talk. By streamlining the group combat experience, Mythic may be shortchanging themselves in terms of potential relationship formation in DAOC.

Interdependencies

Apart from situations where players are already grouped, game design elements can encourage players to interact with each other on a one-on-one basis to differing degrees. EQ, when compared with DAOC, has a system where players are more dependent on each other. For example, a lot of crucial or useful abilities in EverQuest are utility spells that only certain classes can cast on others. Among these are Bind (safespot creation), Resurrection, Clarity (mana regen), Spirit of Wolf (movement enhancement), Teleports, or Invisibility. In DAOC however, Bind is an ability all classes can perform by themselves, cheap public horses take the place of Teleports, Resurrection is a low-level spell that several classes have, and most utility spells can only be cast on the character or on group members.

====


Facilitating Altruism

There are several reasons why player dependency encourages relationship formation. On a superficial level, it increases the possible interactions two players could have. But it’s much more interesting than that. First of all, it increases the number of ways that players can help each other. Very frequently in EQ, you meet someone new by asking for a Bind or a Clarity. The asker is humbled, the giver is empowered, but both players usually come away from the encounter with a sense of mutual benevolence. Asking help from a stranger or being asked a favor from a stranger are far rarer occurrences in DAOC because of the relative independence the game mechanics give each player. These encounters, which are frequent in EQ and rare in DAOC, help create debts of goodwill on an individual level that foster future encounters between these two players. The following account highlights these kinds of relationships:

“My primary character is a Cleric, so on one occasion my guild was on a raid in a dungeon area and I came across one players corpse. This was unusual because of where we were and how deep we were in the dungeon. I sent this person a "tell" to see if she needed a res. She replied and was very excited that I was there to res her. After she gathered her equipment she tried to give me some Platinum pieces, which I refused since I didn’t go out of my way to help her ... I was just there. A month later, my guild was performing another raid and we were wiped out by some unexpected baddies .. The person I ressed happened to be in a group near the beginning of the dungeon where we were wiped out, and before I knew it, most of her guild was there to help clear the dungeon and get our corpses back. I mean about 30 other players went out of their way to come and help my friends out just because I helped one of their friends a month before. I don’t know many people who would do that in real life … All I can say is ... Thank you Ostara” [m, 32]

Random Acts of Kindness

A variation of this theme is the random acts of kindness that many players experience. By increasing the number of ways that players can help each other, it increases the chances that altruistic individuals help lower-level players. Individual altruistic events promote trust at the community level which is crucial for trust at the individual level when two strangers encounter each other and could potentially form a relationship.

“One of my fondest memories of the game was having my first buff cast upon me by a level 19 Shaman. I didn’t realize this could be done and it was at this point that the level of player interaction became apparent. A random act of kindness that one rarely sees in real life these days that has encouraged hours/days of play since.” [m, 25]

“Those random acts of kindness really make online games a pleasure to play in. Whether someone has tossed me a heal, SOW or other useful spell for no reason, or given me a nice item without asking. I've tried my best to return these acts to others whenever possible.” [m, 28]

Some EQ players were vocal about the annoyances of the player dependencies in EQ, and DAOC was consciously designed to make players more independent of each other than in EQ. However, these minor annoyances may actually help encourage and sustain strong social relationships in the long-run.

====


The Mechanics of Death

Beyond specific game mechanics, the world of EQ is also more dangerous than the world of DAOC. In EQ, when you die, your items stay on your corpse and you must travel to your corpse to retrieve your items. There is the chance you may not find your corpse, and also a chance that you may lose all your items if your corpse decays, apart from the frustration of having to retrieve your corpse instead of gaining XP. Both teleports and resurrection can only be cast by one or two classes, so dying is a very “expensive” event in EQ. DAOC is much safer in comparison. Your items stay with you instead of the corpse when you die; everything is a horse-ride away; you can’t de-level because of experience loss; and resurrection is a low-level spell that several classes have. Trust is forged through dangerous and high-risk situations. You don’t ever need to trust anyone except when the situation is dangerous, and EQ does this much better than DAOC. The game design decision to make death easy in DAOC also makes players more nonchalant about dying. Dying is a trivial event in DAOC. But because trusting friendships are forged from dangerous encounters, the mechanics of death actually have a huge influence on how relationships develop.

Conclusion

Of course in listing all these differences between EQ and DAOC, one has to keep in mind that game design is about compromising among multiple objectives, and Mythic purposely chose to streamline certain game features while Verant streamlined others. One might get the sense from the above contrasts that Mythic made poor decisions. This is not meant to be the case at all, and it must be pointed out again that game balancing oftentimes leads to compromises such as the ones mentioned.

In single player and limited multiplayer games, system rules and game mechanics mainly have an impact on how fun and engaging the game is. In MMORPGs, game mechanics have more far-reaching effects. Differences in game mechanics influence how an economy develops as well as how social relationships form. As upcoming MMORPGs provide integrated real-estate and player-elected governments, one could imagine using these worlds as social or political simulations in an attempt to understand large-scale human behavior without the fear of inflicting real world consequences. Or perhaps, we might come to realize that the rules of social interaction in online environments are so different from those in the real world that we need new theories to understand these phenomena.

Questions for Readers: Are there other game mechanics that influence the formation and development of social relationships or social networks? What other interesting large-scale behaviors or phenomena do game mechanics affect? (comment below)

Posted by nyee at 8:44 PM | Comments (23) | TrackBack