 |
Forcing interactionDecember 9th, 2005 |
A comment on a previous post prompted me to dig into an issue that has been tossed around a lot on the blogosphere, most recently in Jason Booth’s blog.
A long time ago now, I wrote in an essay called “On Socialization and Convenience” that
On LegendMUD (a fairly GoP environment, fundamentally) we added a socialization area with a bunch of nifty social facilities. The Wild Boar Tavern offers a lounge for chatting, goofy food to buy, an auditorium, a gift shop to buy goofy items like birthday cards, a wedding shop for in-game events, etc. You can reach it instantly from anywhere by merely typing “OOC.” It was there in an instant for anyone who wanted it.
It doesn’t get used.
On UO we had taverns with NPCs, dart boards, chess boards, backgammon, dice. There were multiple ones in every town. You know as well as I how crowded they were.
Leisure time in a mud is pointless time in players’ eyes, and only a small subset of your players will be looking to spend pointless time. (emphasis added)
The fundamental reasons for this, I believe, are incentive structures and opportunity. In that essay, I wrote extensively about the opportunity issue. In a nutshell, if the action is too fast and furious, people cannot take the time to converse. The faster the pace and the fewer the leisurely moments, the more likely that the socialization will reduce down to basic cues (shout-outs, expressions of fiero, “gg” remarks, disses, and so on). I don’t want to underestimate the important value of these — I just recently read an essay, can’t remember where, in which an academic evaluated their importance actually — but they aren’t the thing that leads towards lasting social relationships.
This led me to say that “socialization requires downtime” — which I didn’t mean as “put lots of tedious stuff in your game” but rather as “think about the quiet moments” or “don’t have a relentless furious pace.” If the cognitive demands on players are continuously high (as in an FPS game, for example), they will focus their attention and concentration, leaving no room to process other sorts of input. Providing moments when the attention can be divided or engage in forms of longer-stage planning is merely to provide an opportunity for the player to multitask. If you picture player attention on a task as a graph of intensity, it would look something like this:
In those games where there is that pace, we find the characteristic that the sessions tend to be short. It’s abnormal for humans to remain at that high state of attention and adrenalin for very long. Indeed, stuff like flow is characterized by high attention but not by high excitement, or else it would be far too exhausting to keep up. In between the short sessions, players do things like watch the remaining players finish up the match, trash talk on chat, or go onto web forums to actually engage in community building.
Essentially, the precept is that you can only put a given amount of cognitive burden on a player, and socializing is a cognitive burden too. If you want socializing, you have to reduce the burden on some other front. And if you don’t reduce the burden, players will do it themselves — they will choose to stop playing for a moment to catch their breath. In other words, you have a choice as to provide rest stops, or else players will pull over by the side of the highway whenever they feel the need to. Thus the creation of social spaces such as the variously located trading spaces that sprung up in EverQuest in different locations over its history, or the UO blacksmith visits, or the example Grimwell cites in Asheron’s Call, the game Jason worked on:
Interesting is the twist that Turbine has experience with ‘hubs’ in AC1 that had almost no decoration. The ‘Hub’ lies West, Northwest of Arwic (which was ruined last time I saw it) and was a nexus to points elsewhere in the world.
The decoration of the hub was minimalist to say the least and yet it turned into a dynamic player gathering point serving the same functions as a town would (rest, grouping, trades, etc.). Because of the convenience in travel that the Hub offered (as a jumping point to multiple other activities, the players chose it as a central nexus. Even without the decor; proving that utility > appearances when it comes to hubs.
So that’s opportunity; but then there’s incentive structures. In the case Grimwell cites, the incentive was convenient travel, which translates crudely into “faster acquisition of experience.” And that’s where the above-referenced comment and Jason Booth’s post come into play.
Have a look at this highly interesting graph from the fascinating PlayOn project at PARC.
As Eric Nickell and Nick Yee write in the accompanying blog post,
Characters on PvP servers level more, spend more time playing, and are fastest at leveling than characters on other servers. Notably, characters on RP servers level the least even though they spend almost as much time playing, but are the slowest levelers.
The conclusion seems obvious — RPers level slower because they spend more time chatting. Since there is only one incentive structure in the game — experience points — players who divide their attention are penalized relative to the average player. In other words that verge on hyperbole, the game’s incentive structure punishes people who socialize.
Now, this is far from the whole picture, of course, because there are other factors that push people together. Most evident among these is forced grouping. The same incentive structure is used in order to convey the clear message that greater return on investment can be had by pooling resources. In other words, two guys killing orcs together can kill them more efficiently.
The thing that I want to emphasize here is that forced grouping does not force socialization, it forces teamwork, and they are not the same thing. Killing orcs is still a peak on the attention graph; nobody is having personal conversations in the midst of an intense raid. It is only because there is the likelihood of downtime in the close proximity of your groupmates immediately before or after the fight that any socialization can occur at all.
Again we can look to the FPS games for the stark example; all the team-based games, such as CounterStrike, put people into groups, but are too intense for socialization during the match. Little personal touches generally come out between matches or while waiting for the match end, when one has the time to type or chance to speak coherently.
So forced grouping is, paradoxically, problematic in terms of socialization because the incentive structure rewards maximizing the XP return during the time you are together, as opposed to maximizing getting to know one another. It pushes people together at peaks in the graph, not at troughs.
When Blake says,
Have the social interactions, of various types, but don’t force a person to have to spend X amount of their time doing that just because you (dev team in this case) think they should. Reward those that enjoy the social interactions, and participate in some way, but don’t penalize those that just want to whack the next shiny mole because they don’t want to spend an hour in a virtual bar watching a spinning character in order to go on with their desired playstyle.
the part that is getting ignored is that the vast majority of players will choose the most obvious path towards cognitive engagement, and towards the reinforcement effects of the reward structure of the game. In other words whacking shiny moles is easier, more predictable, and better rewarded, so lacking opportunity and incentive, people won’t do anything else.
What is the incentive structure that incentivizes forming friendships while in an attention trough?
As an aside: why even care about this? It’s worth following the logic here from a developer’s point of view. Most players come to play these games with already-established friendships these days. That’s a problem from the operator’s point of view because it means that the entire social unit can come and go relatively easily, in terms of its social network; it’s a self-sustaining circle of friends. The hypothesis (and it may well be completely wrong, but it’s based on fairly standard social network theory) is thatthe more tightly the group is “webbed in” to disparate groups in the game, the more likely it is to stick and thus continue to form a part of the community and possible ongoing revenues.
This is where I disagree with several of Jason Booth’s comments. He says, for example,
It’s a fool’s proposition to spend a large portion of your development resources on a space which provides no additive game play time, and generally only fools are willing to push for those types of trade off’s.
The question here is, of course, what is meant by “additive gameplay time.” If it means whacking orcs, then it should be evident from the above that having a space in your game where you don’t have to whack orcs is indeed valuable, even if its only purpose is to allow players to go there when they themselves decide they need a break. Otherwise your game will be go-go all the time, and you’ll be in FPS-land, with an attention graph with no troughs at all.
He also says,
Some designers argue that running from place to place helps chance socialization, but I’d argue that unless you’re the type of person who commonly meets people at the mall, you’ll probably run past everyone and not bother to socialize. Again, it’s the activity which provides the means and reasons for socialization, not the proximity.
But without the proximity, there can be very little opportunity (a bit, true, using tells and the like, but not much). The mall Jason cites is a bad example for a variety of reasons, of course; generally, the walkways of a mall (technically, I suppose that is the “mall” proper) are interstitial spaces; we are on the way somewhere else, and while not at a peak on the attention graph, we’re not at a trough either. We’re somewhere on the upslope, usually, heading somewhere with a modicum of intent.
A better example is a bar, or other such “third places.” The activity there is often ill-defined. They are spaces architected for troughs, rather than peaks. There are activities present, but they are generally intended as low cognitive burden activities. They are designed with lots of downtime in them: bar trivia games, a band with frequent breaks, eating food, drinking alcohol, pool tables.
The PARC guys also did a paper examining the specific case that no doubt both Blake and Jason are referencing, which is of course the use of “battle fatigue” in Star Wars Galaxies. For those who never played, this was a mechanic that accumulated damage to a particular stat that could only be healed in a cantina, which were specifically located in towns. The intent was to provide a designed arc to the attention graph, timing a trough with the natural end of a play session and bringing players to a “third place” at the right time.
When Jason says, “waiting to heal special ‘cantina only’ wounds doesn’t make me want to socialize; it actually puts me in a foul mood” and Blake says “don’t penalize those that just want to whack the next shiny mole” I think what we are seeing is that above all, the designed arc was mistimed. People do not resent even forced downtime when it comes at natural ending points. After a sports match, we are forced to stop, and yet we relax. After a raid, we’re not all rarin’ to go again. The examples are legion. It may well be that the biggest flaw with battle fatigue is that it accumulated too quickly, and didn’t clear itself during logged out time.
If I summarize the empirical data from the paper that the researchers connected to Oldenburg’s concept of “third places,” I get a list like this:
- too many players “hit and run” the cantina, so they never really socialized
- too many players macroed the services offered, and were bots
- this may be because the players present are there not in a social role driven by their personalities — “third places do not set formal criteria of membership and exclusion… the charm and flavor of one’s personality, irrespective of his or her station in life, is what counts”
- cantinas may ironically have been too crowded, or not thorough enough in forming a “mixer” environment
- but cantinas that were not overrun with spammers did have a high “fun” index, showed evidence of socialization and conversations, and so on
- cantinas need “regulars” who provide the soul of the place; by measuring various things, they concluded that the regulars weren’t actually providing it. But many of these regulars were probably spammers and grinders, not true regulars…
In the end, the researchers concluded that “socialization requires downtime” is imperfect — or, as I could phrase it, incomplete. They recommended the addition of things to create more “regulars,” high population density to create an urban feel, better placement of the cantinas so that they weren’t a pitstop, and so on. They also explored the notion that other games, such as City of Heroes which wasn’t yet out, might treat the whole game as the cantina instead; my personal take is that this isn’t really how it panned out.
My take on what happened with SWG’s entertainers: The original intent behind the entertainer profession was to provide an incentive structure to regulars, to get them off of the last spot in that advancement rate graph. Regulars were hoped to provide the entertaining conversation because it was a draw for customers, so the incentive structure still worked. It worked OK, though not great because of the “hit and run” factor, until it intersected with the instrumental play of hologrinders, which resulted in the spammer problem, which then totally undermined the entire concept.
In the end, we’re left with some basic questions and observations:
- Do we really need people to form these new in-game friendships, or are the ones they had before entering the game enough?
- Attention troughs are always forced, either by mechanics or even by plain old player fatigue. Can we architect them, or not?
- Can we overcome the fact that forming new friendships in-game is to a large degree disincentivized by the game systems?
To my mind, the answer to these is far from cut and dried at this point.
You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed. [?]Type in a relevant tag, and click the button, and help organize this blog's information.
[More Help]
).
Bartle?s 5 most important folks in virtual worlds [IMG] Posted by Raph’s Website [HTML][XML][PERM] on Fri, 20 Jul 2007 03:12:38 +0000
[...] Raph has a great post about thoughts and studies on encouraging socialization up on his blog. Go read while I cogitate. [...]
[...] Raph Blogϥǥ˴ؤ¿ǶǤPlayer-centered designפȤǽ SWG ˤĤƸڤƤꡢForcing interactionפȤ SWG ΥƥʡˤĤƸڤƤꤷޤ [...]
Five Reasons Why Web 2.0 Matters Dion Hinchcliffe’s SOA Blog The 46 Best-ever Freeware Utilities Raph’s Website » Forcing interaction NPR : The Boundless Gifts of Richard Pryor Where’s the ‘Wow’? Xbox 360 Has a Better Design, but the Games Fall Short
while innovative, really ought to have just been a ‘free-to-all’ social emote type of thing, as they had a tremendous function in the downtime attention trough parts of the gameplay experience. More on this here, from the horse’s mouth; Raph Koster’s Blog - Forcing Interaction. The article contains some telling self-analysis on what he thinks went a bit wrong with the SWG Cantina system as first implemented by his own team, but also insight into the attention cycles of ordinary MMO Gamers. See? It is bad to be fighting ALL
[...] Popular Posts Do levels suck? (2734)Do levels suck? Part II (1543)Forcing interaction (1099)Where does popularity come from, or the Wisdom of Crowds revisited (852)The future of content (784)From instancing to worldy games (764)Appealing to women (753)The evil we pretend to do (586)Some games worth playing (571)The end of the world (539) [...]
Blogroll Joel on SoftwareRaph Koster Sunny Walker Thoughts for Now Sex, Lies and Advertising