A New Beginning

It’s about time that I dust this blog off and bring it back to life, but with a slightly different focus.  Previously, I was working on a 3D graphical roleplaying world.  I’d gotten my start in virtual worlds in the Neverwinter Nights persistent world community and had set out six years ago on the journey of creating my antidote to the restrictions and magic-circle killing aspects that platform.

The result has been two open source projects.  The first was a domain specific language for defining rulesets that could handle a deep physics model with everything from object assemblies to multiple identities.  The second project was an processing engine for this language.  It was built to scale across multiple cores on multiple machines in clusters and in cloud deployments and I’m certain, that like any professional MMO architect’s first stab at such an architecture, it would cause bemusement among specialists in that field.  Lastly I was going to mate Angela to the game engine Torque 3D.

After 20,000 lines of code sunk into in Angela, I had a functional core of a ruleset processing engine, but had yet to start on the “mating it to an engine” leg of the journey.  I’d estimated that it would take only a couple of months of evenings; which probably means a couple of years.  Only then could I could start world building in earnest.  Yes, I had learned a lot, I mastered new kinds of programming (e.g. functional programming) and produced an interesting, albeit wonky, tool.  I could look forward to administering the ultimate roleplay experience as my retirement hobby.

Then something happened… I joined the rest of the world in becoming the owner of a smartphone.  After owning it for a little while, my thoughts turned to the kinds of worlds that could be built with it; “worlds” that combined aspects of augmented reality, or alternate reality gaming and LARP.  The older posts on this blog are about the old project.  The revival is about that new journey.

Do a google search for “augmented reality” (AR) and the top results are all of the “gee-whiz, this will be big, sometime real soon now” variety.  Wikitude, the flagship AR app is mostly a gimmick, despite being useful on occasion.  Add “MMO” to your search string and you get multiple references to “Parallel Kingdom”; which is basically level grinding meets map app.  I’ve got a game on my phone; a tie in the one of the Paranormal Activity films.  It is no better than one would expect from such an origin; being a Skinner box in a map app, with occasional 3D models superimposed on the camera feed.  Changing the search from MMO to games gives us some links to collections of reviews of gimmicky games.  AR is a medium that has not yet had its conventions defined.  It is wide open and everything lies in front of us; like the web circa 1993 or virtual words themselves around the same time.  AR is still in its cheerleader phase, with people like Robert Rice making the case.  I remember him from Meta Café a few years ago, when he we still working on immobile MMOs, before making the jump to AR.  I don’t know what his ideas for the medium are, but I’m sure he has them.

I have my own ideas for the medium.

I used the imprecise “the medium” for a reason.  If you accept that the term “augmented reality” is only relevant if you are superimposing graphics onto a mobile screen then that has a very specific, niche meaning.   I prefer the term Pervasive Alternate Reality.  It is mobile, inherently multiplayer and may or may not use graphics; but it is always on.  The player never logs in and never logs out.  Physical space matters.  Location matters.  Some of the long established norms of online play get turned on their head when the players come into physical contact.

So now, taking the experiences that I learned building Angela and working on that old school, immobile, world; I’m going to set out to create something entirely new in a new medium.  This could get interesting.

Real Players, IC aliases

My post last week on pseudonymity discussed some of the benefits to immersion that pseudonymity brings, as well as some of the problems for online communities that pseudonymity creates. How might we be able to have the best of both worlds and avoid the worst of both worlds? There is an answer and we need look no further than the Society for Creative Anachronism.

When you register membership in the SCA, or register/check-in at an event, such as the annual Pennsic War, you use your real world identity; what SCA members refer to as a “mundane” identity. Because using your real name can be immersion breaking – especially if it clashes with the period and location of the persona you are undertaking to recreate – people take an alias – known as a “scadian” name – for colloquial usage. The purpose of a scadian name is not to hide your real identity, but to provide an immersion preserving nickname. You can’t escape from the consequences opf your actions. If you annoy others and change your scadian name, you won’t see any difference in how others react to you.

How might this paradigm of using real world identification for administrative purposes and colloquial aliases with persistent reputation work out?

There is an excellent mechanism for using real world identities; Facebook Connect. People use Facebook not as another online only identity, but as a mechanism to keep in touch with their offline friends and family. Because of this, Facebook users typically have large investments in their Facebook identities and users are required to use real identities with Facebook and fake or anonymous accounts are closed. Facebook seems to have cleared the hurdle of users accepting the usage of their real identities, though older users are more likely to still use a semi-anonymous FB profile or stay away from Facebook altogether on privacy grounds.

How might we prevent the immersion breaking scenario of knowing that the elf priestess is being played by a balding man named Frank Smith? Allow Mr. Smith the option of creating and maintaining player aliases; essentially a meta-avatar or a pen name. Alternatively, every character could also stand in as a player alias. These aliases are for IC use. Players can use a nickname when making OOC posts on the world’s forums, but they can’t use the IC aliases to make OOC posts. Conversely, IC posts on the forums can only be made with the IC aliases and the only in-game identifier is the IC alias. Players can choose to show or hide their alias links when their profile is viewed by other players. GM staff and admins can always see the linkage.

Establish a player karma system similar to Slashdot there the dimension being measured is the congeniality of the player; on the scale from total jerk to pleasant to be around. This measure should not include roleplaying skill, but merely how much fun this individual is to play with. A player may have only give rating per alias that they encounter, though they may change their rating at any time. These ratings are collected on the user and not kept separately and the rating of the player is visible to anyone. This means that a player who consistently acts like a jerk will show a poor rating with a new alias, while a player who is pleasant to play with would show a good rating with their new alias. This way, an established player with a good reputation may also use that reputation anonymously. Other players may not know who is playing the new elf priestess, but they know that she is being played by an established player with a good reputation.

The cost of pseudonymity

Nearly every roleplay environment that I’ve ever seen is Pseudonymous; meaning that the players use pseudonyms and are essentially anonymous. Anonymity can be liberating. Some players regularly create alternate player accounts to avoid metagaming by other players. It is often informative to see the difference in how certain players treat your established personality versus the anonymous one. Some simple do it to have a quiet chat panel. Some do it to firewall in-world identities from offline ones; whether it is the man who does not want his friends to see how often he role plays female characters or female players who want to avoid sexual harassment. Lastly, knowing that the elf priestess is played by a 40 year guy from New Jersey who works as a tattoo artist and whose other hobby is his Harley might be just a tad anti-immersive.

For the GM team, the ability to play incognito is a virtual Swiss army knife. GMs also like to play and may simply want to relax and be a player for a bit; free from a barrage of GM related requests from players. They may want to avoid players acting differently because they are in the presence of a GM. They may use it for research purposes. A “newb nobody” player is more likely to see open cheating or griefing than a known GM moniker. Lastly, they may use it for “NPCs in PCs clothing”. Players often mentally separate PCs from NPCs; especially when the difference is obvious, or there are means of doing so (such as a currently active player or character list). The GM team can use a player account to give an NPC a PC cover for certain uses.

Anonymity has a well known downside however:

We’ve seen it all; the griefers, the problem players, the trolls and the cheaters. Everyone has seen the banned player pop back up under a new moniker, a new IP address and a new CD key (where appropriate). Meta roleplayers may be harmless to a community, or even beneficial in preserving immersion; but if they are highly networked, as in the case of “Karyn” of LegendMUD (probably the most famous example), they can produce a “community bubble”. A community bubble is a short term strengthening of an online community at the cost of its long term cohesiveness; often induced by highly networked meta-roleplayers. Previously, I was somewhat neutral on the subject, but I now wonder what effect that the player I wrote about earlier had in the ultimate demise of “her” particular online community. This is not unique to online identities, but it certainly is more common online than offline. A griefer inspired one of the earliest media articles on virtual worlds, “A Rape in Cyberspace”, before most had even ventured online. That banal incident of greifing on LambdaMOO would be a dog bites man story today; there was not even any corpse camping or teabagging involved. In the early days of public awareness of the internet, it was still newsworthy. Mr. Bungle’s account was closed by community vote afterwards, but it was likely that his actual usage of LambdaMOO was uninterrupted as he could simply come back under another moniker.

The Karyns and Mr. Bungles impose a definite cost on community cohesion. Friedman and Resnick’s paper, “The Social Cost of Cheap Pseudonyms” (warning, 20+ pages of wonky, academic discussion) explores low cost Pseudonyms as a game theory problem. The conclusion that they reach is that the most effective – but still highly inefficient – limiter of bad behavior under the cover of pseudonyms is treating all newcomers to a community badly until they have proven themselves. Unproven Ebay IDs often don’t have their purchases shipped until after the check clears, while users with high ratings get their shipped right away. Many forums don’t allow PMs until they have a certain number of posts. Friedman and Resnick do agree in their paper that the ideal should be to treat newcomers well until they have shown that they should be treated badly, but that low cost pseudonyms decrease the cost of defection too much. “Defection” is game theory speak for “being an ass” and game theory is a branch of mathematics exploring that part of human psychology related to putting the screws to our fellow human beings.

The simplest example of game theory is the prisoner’s dilemma. If the prisoner’s dilemma, the participant (known as a player) is told that he has been arrested and the police have enough evidence to convict him. He is offered a deal. He can testify against his cohort, or remain silent. If he testifies against his cohort (the defection choice) and the other remains silent, then he walks free and the other fellow spends 5 years in jail. If he is mum and the other fellow testifies, the reverse happens. If both testify against each other, then both server 4 years and if both are mum, then both server 2 years.

• Obviously, the best outcome for the player is if he defects and the other is quiet.
• If he feels that he can’t trust his cohort, then he must defect to minimize the damage.

One critical aspect of the game is whether it is played as a one off affair, or has subsequent rounds (known as the iterative prisoner’s dilemma). When played over many rounds, a player is subject to retaliation for earlier defections. The most common strategy in multi-round games is known as “tit for tat”, where a player does not initially defect (i.e. they remain silent) but subsequently repeats the last action of the other player. This threat of subsequent retaliatory defection is a very effective deterrent against defection. When the game has only a single round, defection is clearly the best choice.

Low cost pseudonyms essentially allow one player to play one off PD, while the other is playing iterative PD. Because of this, long time personas have to assume that any new persona will defect; making the community less welcoming to newcomers. Since online communities always have a certain degree of turnover, they need new members to replace retired or inactive ones and a less welcoming community has to work at a handicap to stay healthy.

Next time, we’ll explore a potential strategy for handling player pseudonymity.

Ad Hoc roleplay groups in the great desert

Jim Moreno just posted an article about forming ad hoc roleplay groups in WoW. He calls them roleplay troupes and contrasts them with guilds. The advice could fit any similar theme park world, or any any commercial AAA MMO environment for that matter. I generally favor building a small RP focused world using whatever platform/engine as opposed to trying to find water in the desert; but I do know some folks who do it and this article has some useful tips for them.

Trope Cloud

I made a discovery a couple of weeks ago; one that endangered my work productivity for a couple of hours (don’t worry employer, I made it up and made my deadlines). There was a thread on the Amia forums about playing evil (the one that prompted my post on playing evil in fact) and someone linked to this article as a tip. This set me to exploring the TV Tropes wiki. What an absolute goldmine!

It also gave me an idea about incorporating feedback from other players into our playstyle.

Take any character from any world and consider just how many clichés belong to that character. Sure, that character may be anti-hero number ten thousandth and one on the server and no, I’ve seen the crazy gnome wizard cliché before. I’ve played the crazy gnome wizard. And the mentally unstable Viking. And the goody two shoes paladin. And the paladin dressed in black (ohhh the edgy combination of paladin and black. We’re almost in Batman territory now). And the wolf-in-sheep’s clothing. I’ve lost count of the number of Longstocking clones I’ve played. Etc. Etc. It is very likely that another player sees your character and thinks “here we go again”. But there is an important element of the equation here; the player may be exploring the dark anti-hero trope (or any other) for the first time. Never fear. Just about every story in existence can be broken down into a list of clichés (tropes), and its not as if the roleplayer is doing something that every novelist and screenplay writer is also doing; even if by accident.

This got me to thinking about how we choose our characters tropes, how we play them and how we get feedback from other players. One thing that I’ve notices is that there is something of an unwritten rule that at any given point, the main characters of the playerbase will cover a broad spectrum. They won’t all cluster in anti-hero, villain, hero, rogue, mage, etc. tropes at the same time because too many characters of the same type limits the individuality of any particular character. Character X may be the nth paladin in the history of the server, but one of only a handful of currently active ones. Chances are, the player does not want to be “mysterious, yet unmistakably powerful, good, drow #167”. They may not even be aware that their character is perceived differently than they intend.

Enter the trope cloud. The trope cloud is a tag cloud, but the tags are all tropes. When playing, other players could semi-anonymously trope tag a character. When reviewing the character sheet, the player would see the trope cloud as they would a tag cloud (such as the one in the right side margin of this blog). The trope tag would be semi-anonymous because the player would not be able to see who tagged there character with which tropes, this information would be available to the GM team as a safeguard against abuse. Clicking on a trope tag would bring up a list of other characters tagged with that trope. The global trope tag cloud could also be viewed by a player considering making a new character, or planning where to go with a character. Hopefully, it could help them avoid oversubscribed tropes.

Plus, it might be fun to hang lampshades on our characters on occasion. As long as it does not get out of hand and break the immersion of the world.

Breaking Combat Addiction – at least certain kinds of it

Flatfingers had a response to Brian Green’s question “Do you enjoy your favorite MMORPG more or less because of the changes that have been applied to it?”
He answered:

“Unhappily, my MMORPG experiences since EQ have led me to precisely the opposite conclusion: as a gamer, I’m just not interested in playing any of these games any more because my perception is that they have ceased to change in any meaningful way.”

I originally had a draft post that used the discussion on live team drift to lead into Flatfingers post because it fits with the question I raised last week about murder and theft as core gameplay; whether it is possible to get away from it. Flatfingers is correct when he laments the “lack of innovation” in the MMO industry. The typical MMO is “kill and loot” + agro + trinity. Even nontraditional, child oriented MMOs, such as Wizard 101, follow the same formulas. This is not just with the online RPGs. Change the ruleset and tweak the story (but keep the cliches) and Dragon Age is another Baldur’s Gate. Pen and paper RPGs have been in the “commit mass murder and steal from the dead” business ever since at least 1978. Aggro came from text MUDs and was initially a workaround for the problem of determining who a mob attacks in a node based environment that lacks a coordinate system.

But what could come in its place? Some worlds do break from the “standard model” in spectacular ways:

• RPI MUDS
• A Tale in the Desert’s crafting emphasis – not a dead mob in sight
• Darkfall’s manual swings
• EVE Online’s time (as in calandar time) based character advancement and economy emphasis

It is probably not surprising that the first two of these examples are socializer and roleplayer oriented, while the latter two are PvP oriented. But what about for achiever and explorer oriented play? As Damian Schubert has often pointed out, the usual gameplay conventions are not in place because of lack of innovative thought, but because most of the alternatives simply can’t hold onto a player for an extended period of time. As evizaer pointed out, breaking combat addiction is hard to do, very hard.

One problem with player pontification is complaining that there is no innovation without trying to come up with alternatives. We can’t just expect professionals to “innovate”. They are usually spending other people’s money and have to take a route that is known to “be fun”. We have to come up with alternatives and ideally implement them; or at least prototype them. My conjecturing from yesterday is worthless without a prototype example.

So I repeat my question from last week. Is it possible to design a world that is compelling to explorers and achievers without resorting to the usual crutches?

Live Team Drift

Last week, Brian Green wrote about how individual worlds change over time, specifically using WoW and Meridian 59 as examples about how various rebalancing and bugfixes change the dynamics of the world (e.g. with ranged weapons in M59) and how demographic changes in a world’s player population prompt changes as well (such as the erosion of immersion in the Wolfshead article that prompted Green’s post). As far as I know, there is no specific term for this, though “Live Team Drift” might be appropriate. Community worlds are just as prone to this as commercial ones. Amia’s recent introduction of a “job system” is a specific example. Thrym, the lead admin of Markshire and one of IGN’s Neverwinter vault editors, constantly introduced new subsystems to Markshire and while I was the lead dev on Etillica, I was also prone to major overhauls of the world.

The reasons that live teams do this are many. If fact, community worlds are probably even more prone to it as they tend to launch in less than polished states (pre-alpha being more the norm) and are constant works in progress. Sometimes it is to fix glaring bugs or imbalances. The list of NWN worlds that started out with a “we don’t want to nerf or change spells” policy and then went on to customize (and nerf) the scripts for nearly the entire spellbook is long and illustrious. Sometimes, it is to try and make the world truer to its ideals for its actual (or intended) playerbase; such as devdisco’s job system on Amia. Sometimes it is to expand or deepen a particular gameplay aspect that the players or admins are fond of; such Hephaestus’ extensive additions and modifications to Mythos’ implementation of Craftable Natural Resources (CNR). The fact is that any actively administered and developed world will change over time and possible even be unrecognizable to earlier players.

Green asked the question “Do you enjoy your favorite MMORPG more or less because of the changes that have been applied to it?”

This is like asking about your favorite model year of your favorite model of car. Is the latest Ford Mustang the best? One from recent years? Another from the late 60’s? You’ll never get an objective or consistent response.

Train to Fight, not Fight to Train

Let’s undertake a design exercise. Our constraint is that we are trying to come up with gameplay that does not rest on mass murder. Yes, I wrote “gameplay”. Even in hardcore RP worlds, you need gameplay if you don’t want the world to simply be a chatroom. Roleplayers also have achiever and explorer motivations (when using Bartle types). For our exercise, we’ll use a mêlée combat system in a standard fantasy world. It does not matter is it is high or low fantasy, though it may be a better fit for low fantasy. Similar thought experiments could be carried out for magic, crafting, economic, etc. systems, but we’ll focus on simple mêlée combat for this post.

For some of the core ingredients of our recipe, let’s look back to the deep physics series.

• We’ll approximate a character and its equipped items as a cloud of point masses. This includes things such as armor and weapons and can also include the physical build of the character; tall versus short, stocky versus slender, etc.
• For each animation, we can then calculate the transient moment of inertia of the character for each keyframe of each animation, which we’ll call the Keyframe Moment of Inertia, or KMI for short. Each time a character changes his/her equipped items, we’ll need to recalculate the KMIs for all of his/her animations. We’ll keep a remote process that can calculate the KMIs of all animations for a character, sending the current point mass cloud as input.
• Once we have these KMis, we can take the strength and skill of the character and determine the new timescale of each keyframe. Strong characters accelerate through their animations faster than weak ones. Small and slender characters also accelerate faster than larger ones, but can’t achieve as high a final velocity for heavy weapons. Etc.
• Let’s toss in some random tweaks to the keyframes to account for individual styles.

Now we’ll add a few gameplay bits that are not directly drawn from the deep physics concept:

• Let’s allow players to create macros, or scripts in the form of Memotica Action Choreographies. A chirography is a set of action keyframes (Memotica Action keyframes correspond 1:1 to animation keyframes) strung together. Choreographies can also include if/else style logic and can be nested inside other choreographies. In gameplay terms, this allows player scripting of complex combat logic in what would be a high twitch without the need for player twitch. This also allows the player to create new combat moves and even entire fighting styles.
• Let’s allow characters to analyze each other’s moves and to learn from them or teach others. If a character if performing a particular move in a less than optimal way, another character can take note of this and show them the “right” way.
• Allow a “holding pattern” action that can be terminated with one of several options, depending on what the character sees. You might know this as a defensive stance. If a character sees an incoming oberhau, perhaps he responds with a pflug.

So what does all of this give us?

Well, for starts, we’ve just made the combat system insanely complex; even with the simple act of swinging a sword. Being complex is dangerous and potentially game breaking, but can also be a good thing. Chess is complex. The rules are simple enough to teach a six year old, but it takes years of study to truly master the complex game that emerges from those rules. That is what we are trying for here. What armor you are wearing matters. The type of weapon (or even who made it, but that is another discussion entirely) matters. How you hold it matters. Most swords allow for some adjustment up or down of the character’s grip. Your character’s physical (as opposed to character sheet) build matters. Who you have practiced with and how much you have practiced matters. All of these things are intuitive and fit our sense of naive physics; making them easy to learn and understand, but mastery does not come easily.

Naturally, feints will have to be part of the ecosystem. Perhaps that oberhau is not an oberhau at all, but a feint explicitly staged to provoke a pflug, while the real strike goes to the thigh. How often has the character ever seen that feint, if ever? Can they tell the difference between the feint and the real thing? This means that how much sparring the character has participated and observed matters.
What do we gain from making the combat system so complex? We gain four things.

1. We can switch from a fight to train (via combat XP) to a train to fight model without losing gameplay. Mastering the three dimensional chess that is swordsmanship becomes a gameplay end unto itself. We can drop most or all XP from combat as it has become a way to demonstrate prowess, rather than a way to gain it. Since we no longer have to gain directly from combat, we can make it less frequent and we are free to make it more dangerous as characters can flee or avoid combat without incurring opportunity cost.
2. The characters’ domain knowledge and the players’ domain knowledge converge. This increases immersion and leaves us with fewer OOC ways to break the magic circle. In fact, the player and character learn the ropes together, proceeding from novice to master.
3. Dojo RP – Rather than highly generic and nonspecific “I am a warrior” roleplay, where the player roleplays the style of a knight, or barbarian, or pirate, etc. without delving too deeply into the details, we can have heated discussions of the merits, or lack thereof, of various moves and fighting styles. “Your swing is sloppy and slow there and you are telegraphing yourself. Watch me and I’ll show you how to do it right” and “Be careful when fighting orcs. They are fond of this kind of feint…”.
4. A vehicle for IC character created content where characters can leave their mark on the world. E.g. invent a new school of swordsmanship and be its founding master.

More on Williams’ roleplayer survey

Raph Koster has a new piece on the Williams paper that I mentioned last week, where he hits on Williams’ points. One of the early commentators on the thread was asking about the relative RPer populations of EQ2 versus SWG, presuming that the latter had a higher RP’er population. Actually, I’d take this a step further do a sampling across worlds and technology platforms. It is not just SWG, the Matrix Online and (formerly) Ryzom likely to have higher roleplayer counts. Roleplayers who try to RP in large, commercial AAA environments not only have to deal with having their immersion broken, but are often actively persecuted. You have three choices in such an environment:

1 – find a roleplay guild, sequester yourselves as much as possible and deal with the occasional persecution.

2 – give up on roleplay

3 – leave for greener pastures

IC enforced roleplay worlds one of the last bastions of the NWN persistent world community and is definitely well above 5% of the total NWN PW player population. This is despite the fact that NWN was designed for cooperative multiplay and not PW use and NWN2 was optimized for single player; resulting in a situation where roleplayers have to work against the platform and often players are forced into a form of metagaming and collectively ignore some of the pain points where the diku-like engine is breaking immersion. They put up with it because they can be gatekeepers to their ivory tower and there is no other 3D environment that readily allows roleplayers to play in worlds build and run by other roleplayers.

Text MUDs also likely have a much higher roleplayer representation for largely the same reasons. Roleplayers on text MUDS are trading the 3D environment for more codebase control.

Then again, the total populations of these latter two entries number in the low thousands, probably under comfortably under five digits. Considering that commercial worlds have, collectively, close to 50 million active subscriptions, this is a not likely to affect the total percentages so much.

Interesting Paper – Behind the Avatar

Dmitri Williams of USC recently got his paper “Behind the Avatar: The Patterns, Practices and Functions of Role Playing in MMOs.” and linked to it from a Terra Nova post. It appears to build on Nick Yee’s earlier observations about roleplayers and the immersion motive. The paper is a 30 page word doc and as a warning, it is a social sciences academic paper; which means a low information to text ratio (disclaimer: my academic background is in the physical sciences, where the style is very different). I’m also skeptical of the choice of world; Everquest 2. Who in their right mind will even attempt to RP on a mass market, AAA diku? Yes, I do know that there are RP guilds in WoW and that some people I played with back in my NWN days are now in WoW RP guilds. However, it appears that Dr. Williams was able to find enough roleplayers to do meaningful research and the paper is a fascinating read I and I’d put it – along with Yee’s original Daedalus Project research into roleplayers – into the must read category for anyone building or running an RP world.

Most of Dr. William’s observations are consistent with what most RPers already know anecdotally.

The results suggested that role players are a relatively small fraction of the game world’s population, and that they skew younger and more female than the general EQII population. Role players also tend to come from marginalized offline groups and to have a disproportionately high level of psychosocial and health problems. They appear to role play more to express their true, often suppressed, identities than to negotiate new ones. In keeping with their desire for immersion, they use voice communication less than others. On closer examination, these players also have a rich social fabric in which they display significant creativity. Role players use their spaces as a therapeutic release from their daily lives, and often build genuine communities. Despite this and despite the choice of studied world being EQ2

The hints of lower mental health give me pause as that confirms a negative stereotype that roleplayers have. As far as I’m aware, I’m perfectly mentally healthy. However, twenty years ago, I was a shy, socially awkward teenager for whom reading science fiction and creating DnD settings was a form of escapism. It would be interesting to see how the mental health indicators compare with age within the roleplaying population. Also, I’d be interested in seeing the differences between self identified dramists and simulationists; both of whom regard immersion as a motivator and would be drawn to roleplay.

I have only one quibble that seems to have been missed by the peer reviewers. One of the statistical observations that Williams makes is that roleplayers spend less time logged into the game than non-roleplayers. However, in his discussion section, he adds:

By spending less time in offline social life, these players gain acceptance amongst each other, but perhaps at the cost of integration into the larger society. In turn, social diversity may suffer if these groups leave the mainstream.

This passage suggests that roleplayers are in fact spending less time sequestered from offline society than their mainstream peers. It could also be that roleplayers are sequestered from offline society more than average, but not as much as the non-roleplayers. It is minor, but it is an unclear issue.