The D&D 3e design includes a complex character creation system and encourages players to think in terms of “builds” of which classes, prestige classes, feats, etc., they want to use to achieve their character concepts. These choices are rarely made of discrete elements but are strongly coupled together: Prestige classes have prerequisites, you have to take lower level class features before the upper level ones, you accept the “flavor” consequences of some choices for desired mechanical effects, you accept the mechanical consequences for desired “flavor” choices. Players are encouraged to take these choices seriously, to think about them ahead of time. This encourages commitment to a character concept.

Many elements of the 3e system were widely criticized. One criticism was the need for “system mastery” to know that some of the choices were just categorically worse than others. Another criticism was that many choices included tradeoffs between combat effectiveness and effectiveness in other arenas which made it harder for different characters to harmoniously function “as a party” while interacting with different situations: player A was twiddling his thumbs during the fight, player B wandered off to get a snack while the rest of the party negotiated with the baron. The 4e designers tried to address these issues by making it hard to make bad character building decisions. While the optimizers can still find ways to make smart decisions pay off, the difference between a mechanically optimized character and a non-optimal character weren’t as dramatic. Additionally, the designers made sure that each character had a combat role to play: each character has an array powers that do damage and inflict status effects on monsters. While there are many reasons to commend these design decisions, one consequence is that the choices matter less. This is the source of complaints about the “same-y” feeling about 4e classes and powers. If one mechanical choice is roughly the same as any other then making those choices creates less investment in the choices the players actually make. Less investment in a character could easily manifest as less “roleplaying” and a greater lapse into purely mechanical talk. Since 4e’s mechanical design is more robust than 3e’s and it gives players lots of fun-looking mechanical buttons to push it’s easier to switch into purely mechanical talk and not contribute to rich and robust fiction. As a result we can see that there is some merit to the concern about there being “less roleplaying” in 4e: roleplaying isn’t outlawed in the game, but one of the psychological mechanisms that supports investment in a character is weaker in 4e than the earlier edition.

If we’re trying to extract lessons from this analysis to use in other game designs, I wouldn’t recommend inserting “system mastery” features like feats that only newbies use, but I would recommend making sure that the mechanical choices feel sufficiently different to players that they matter. One way to do that is to make it harder to evaluate choices on a single axis, such as damage-per-round. By making hit point damage the common thread in all 4e powers it encourages players to think about their choices in those terms. On top of that, damage and hit points, as bland totalizing mechanics, tend to blur differences rather than making them more distinct.

In written fiction there are a few commonly used viewpoints: First-Person, Third-Person-Omniscient, and Third-Person-Limited. The first-person viewpoint involves a narrator telling a story they participated in to the reader. This viewpoint lets the author communicate the thoughts and feelings of the main character directly to the reader, and also allows the writer to use the descriptions and narration of the text to contribute to the main character’s characterization since what the character chooses to relate to the reader tells us important things about the character. The third-person-omniscient viewpoint, by contrast, has access to the inner thoughts of all of the characters and the narrator is generally not a defined character in the story. While this allows the author to communicate lots of information to the reader it usually means that only the actual events of the story, rather than how they are related, can contribute to an understanding of character because it usually requires the narrator to be somewhat neutral and distant. The third-person-limited viewpoint is a sort of “best of both worlds” approach. Instead of having access to the inner thoughts of all characters it limits itself to the thoughts of one character at a time. Since the text is so tightly tied to a particular character the author can use description, etc., to contribute to characterization. In some ways it allows the reader to be even closer to the character than first person because there is no particular narrator relating the story to the reader: Harry Dresden isn’t telling me this story, I’m experiencing the story the way Ned Stark did. In addition to those viewpoints, there is also what is sometimes called the third-person-cinematic viewpoint. This viewpoint doesn’t relate any of the inner thoughts of any characters, it only tells the reader what they’d be able to observe if they were watching the scene as they would in a movie. Unless the characters explicitly verbalize their thoughts, the reader can only infer what they are thinking via their actions. This viewpoint has many of the limitations of third-person-omniscient but few of the advantages so it’s rarely used in prose fiction, but it’s the way that most movies or plays would be written.

Thinking about these viewpoints in an RPG context gets very fuzzy very quickly. Looking at the classical player/GM split, the common case is that each player controls a single character so the temptation might be to say that they experience that character’s story in the first-person viewpoint, but I’m not sure that’s true. Even though the player is responsible for deciding the inner thoughts of the character, it isn’t always common to articulate those thoughts. Putting aside for a moment the issue that the player is contributing to the creation of the fiction, would the player’s experience of the character’s story be appreciably different from a reader’s experience of a third person account of it? Since the GM in this scenario tends to not have input into the character’s voice, the biased descriptions that we usually associate with first-person or third-person-limited aren’t used. Since the GM in this scenario is most commonly responsible for describing an environment that any person there would perceive, I think the closest match is actually to the third-person-cinematic viewpoint. Sure, RPG players sometimes articulate the inner thoughts of their characters, but frequently only to the extent of communicating attitude or emotion (“I’m going along with this, but I’m suspicious”, “I’m being nice to his face, but I actually hate this guy”) the way an actor would use nonverbal communication in a film. I think this helps explain why references to film or television techniques, such as describing what the camera sees, are so common in RPGs even though, as a verbal medium, one might initially assume they had more in common with written fiction.

However, in the podcast Todd did offer one interesting example of something that seems closer to third-person-limited: in Apocalypse World, the GM is instructed to bring fictional details to the attention of particular players. By threatening an NPC that a player cares about and bringing that to that player’s attention the GM is subtly shading the world description: you notice this because you are the type of person that would care about this. I think I also stumbled across a technique for doing third-person-limited in game design with my Ronnies game Brick & Mortar: Last of the Independents. Since it’s a GM-less game, instead of relying on a GM to contribute external fictional details I have Apocalypse World-style moves that tell players when to do certain things. By framing the “when” condition of the move in a particular way a game designer can shade how the player will interpret the fiction contributed by the other players. For example, my Survivor character has this move:

Quiet! Did you hear that?: Whenever people are engaging in unproductive bickering, mention the ominous thing you just noticed.

This contributes to the characterization of the Survivor: they are the kind of person that expects the people around them to engage in unproductive bickering. When a player is being vigilant for “unproductive bickering” their perceptions are subtly shifted, similar to how a third-person-limited author would use narration and description to color the story in a characterful way. Although Brick & Mortar has some issues as a game, I think this technique is an interesting one that could use further development. It should be especially useful in the “tightly focused situation with defined characters” genre of games.

• It must represent the backstory events in the character creation mini-game
• It should be able to be usefully invoked to do the sort of things the character is good at
• It should offer good, characterful opportunities for GM compels
• Reading it should give someone a good idea about what the character is about
• It should be “interesting”, punchy, and well-written

Add to this the fact that players will often have only a limited number of aspects to describe a character that might be deeply nuanced, which urges players to use aspects to inform multiple character traits simultaneously. Trying to satisfy all of these goals at once is challenging, even for a good writer. Often, trying to do so many things at once leads to lengthy baroque aspects that are clunky to use in play.

Successfully solving such a writing challenge also tends to lead players to be deeply invested in the particular wording they’ve selected. In prose writing there’s a piece of advice known as “kill your darlings”. The basic idea is that it’s very easy for a writer to fall in love with a particular turn of phrase, character quirk, or other minor aspect of their writing that ends up detracting from the impact of the overall work, even if it’s beautiful on its own. Writers need to be reminded to be ruthless while editing to keep from falling into this trap. All of the pressures on FATE players to write “good” aspects can easily turn each one into a “darling” that they’d have difficulty killing. In the Dresden Files version of FATE, players are frequently given the reward of being able to change an aspect to let their characters change based on the events of the story. In my play of the game this option seemed to be rarely used. Even when I felt that the events of the story warranted changing my character I was reluctant to alter my aspects because I didn’t want to lose the many things that each one seemed to be accomplishing. There were aspects that I found difficult to use in play that I was reluctant to change because they were the only link I had to backstory or character points that I cared deeply about. There were aspects that required lots of uncomfortable jawboning to use during play because they were too cleverly written to satisfy the goal of making double-ended aspects that could be both good and bad. While aspects seem appealing at first glance, I think they are problematic as implemented in FATE 3.0 games because they are trying to serve so many masters simultaneously (and that’s without even going too deep into the tricky topic of compels).

My Fighter Seems Overpowered

The Fighter’s base damage is 1d10. On my first level up, I took the Merciless advanced move, giving me +1d4 damage. On my second level up I took Scent of Blood for +2 damage any time I attack the same enemy in subsequent rounds. If I’m supposed to have +1 damage for wielding a sword like the equipment page says, it would be even more damage (it’s unclear to me if the Fighter’s signature weapon inherits the tags from the base weapon — it would be nice to have that clarified). I have a lot of armor and hit points, so the weak hit and strong hit are both pretty good for me, and I started with a +2 mod on my strength so (if I’m doing the math right) I get a 7-9 weak hit 41.66% of the time and a 10+ strong hit 41.66% of the time. That means 83% of the time I’m whomping monsters pretty badly, frequently killing even impressive beasties in one or two hits. Compare that to a typical cleric, who’s doing 1d6 damage, might have a +1 strength mod, and likely has less armor and hit points (which means the weak hit isn’t a very desirable outcome). He’s getting a strong hit 27.78% of the time, at which point he’ll generally do less than half the damage my Fighter will, and he’ll get a weak hit 44.4% of the time, which will probably hurt pretty badly in exchange for doing a comparatively small amount of damage. In play, it sort of feels like the other player’s Cleric is kind of a spectator in these fights (he’s also a level behind, experience-wise, exaggerating the effects, and I increased my strength at third level so my hit rate is even better). His healing has been important, but I feel guilty for seeming to be so much more effective in combat. Now maybe this is the way it’s supposed to work, and I’m just very effective in my spotlight area while he’d be able to shine when we’re doing something else, but I’m not sure I’ve seen that in play (and I noticed a similar Fighter + other guys pattern in The Walking Eye DW APs). I thought maybe the Cleric would have some sort of edge when dealing with mystical stuff, but he doesn’t appear to.

Having Strength Highlighted Is Boring

This session my Strength got highlighted for experience again, and I noticed I had a lot more fun when my Strength wasn’t highlighted last session because it encouraged me do to a bunch of crazy stuff. Since my Fighter does so much damage, the Hack & Slash move starts to feel like an “I win” button, and the reward for hitting it over and over again starts to feel pretty hollow. While I’m having fun stomping on monsters, I think I have a lot more fun when the fights seem frantic and on the edge of control, where I’m Defying Danger quite a bit and the fight is dynamic and interesting rather than a slugfest.

The Classic D&D Stats Don’t Translate to Enough Stuff in the Mechanics

I gave my Fighter a high intelligence because I thought it would fit the character. The only thing I can really do with that in play is Spout Lore, which I’m finding that I don’t really enjoy. It always feels a little bit awkward when I do it, like I’m fishing for information, and the stuff that the GM tells me on a hit frequently feels like stuff I had already worked out from context anyway, so it ends up seeming cheesy and redundant. I suppose I could use my multi-class move to pick up the wizard move that would let me Discern Realities with Int rather than Wis, but that seems like a weak use of the multi-class move when it could give me something as impressive as spellcasting. Basically, I feel like I’ve made decisions based on the fiction (what my vision of the character is) that are obviously mechanically suboptimal, and that’s not a good feeling. I think that some of the stats are too narrow in what they affect mechanically, such that it’s hard to rationalize spending points on them.

The Aid/Interfere Move Kinda Sucks

Since we’re playing with only two players and a GM I think we frequently forget about the Bond mechanic. We were in a situation last session, though, where it made fictional sense to use it, so I rolled Aid, and then realized that the thing only goes off on a 10+, which means that more often than not you’ll end up making things worse when you try to help. I probably won’t try to use this move again since it seems like such a bad deal mechanically. I don’t like that I’m feeling encouraged to completely ignore this subsystem. [edit: In the comments, Jesse suggests that I'm not interpreting the 7-9 result for this move correctly, which is certainly possible]

Grundloch’s Magic Is Annoying

Dealing with Grundloch’s magic in the intro adventure often makes me feel like I’m standing around like a chump, which is completely at odds with how I want to view my character. Maybe the way our GM is running things has something to do with it, but I think the mind control and illusion stuff that my character is being subjected to is really frustrating, and I can never actually get my hands on the caster because he seems to act via illusionary duplicates. In the last session we were in a room with illusionary monsters that were inflicting a mental effect on us. We knew exactly what was going on, but it didn’t seem like there was anything we could do, mechanically or in the fiction, to make it stop. His ability to arbitrarily act at a distance is just lame, as far as I’m concerned, because all we really seem to be able to do about it is wait until it’s over.

I’m Not Sure What D&D Tropes This Game Is Trying to Embrace

When I created my character, I decided to make him a weathered mercenary type, so his priority in going to the dungeon was to find some treasure, but in our initial explorations there barely seemed to be any, but the “dungeon” included a cavern large enough to house multiple armies. Now our GM is really pressing the “if you don’t stop Grundloch bad stuff will happen!” thing, and it feels like the forced urgency of that is totally overshadowing anything else that I might be interested in doing in the dungeon. I’m not really sure if DW is supposed to be about old-school dungeoneering or whether it’s supposed to be more of an “experience this thrilling action tale across a sequence of set-pieces” thing. Personally I’ve never played any version of D&D on the tabletop (which also means a lot of the lame D&D-isms in this game like the dramatic change in survivability between level 1 and 2 just feel lame to me rather than nostalgic), but my mental image of exploring a dungeon doesn’t involve a numberless horde of baddies or being taunted by a big bad. I think I want more mirrors and ten-foot poles than I’m getting.

As is my tendency lately, I immediately started mapping the ideas in the book to the field of roleplaying game design. The most obvious thing here is that many of the factors that would seem to feed into an overall “fun session” are strongly subject to the Halo Effect. Things like “a chance to hang out with friends with the same geeky hobbies”, “fictional content I enjoy”, and “good game mechanics” all seem like they ought to feed into “fun session”, so if you have a fun session you are likely to remember the game mechanics as being good even if they weren’t (I listen to a lot of AP podcasts, and it still amazes me how often people will say that they enjoy a game after I’ve listened to an entire session of them struggling with it). This is yet another reason that designers need to observe objectively when playtesting to be able to identify what truly does or doesn’t work since self-reporting will be strongly influenced by the Halo Effect. (It also means that people with an established fanbase can produce poor games and have them called “good” by their fans, which is more common than I would like the RPG hobby).

I think it also ought to raise some questions about the “playtest at conventions” strategy that some indie designers have committed to over the past few years. Does the high-energy environment of a con cloud useful data? Does the weird microcelebrity “I played with the designer!” thing pollute your results? I’ve never been to a con so I’ve been reluctant to share my thoughts about this stuff, but I’ve never been comfortable with the idea of mixing advocacy and testing, which seems to be a lot of what the “playtest at cons” movement is about.

Changes in 0.71:

• Changed “Starting a New Chapter” to “Starting Chapters”
• Changed “Playing Out a Chapter” to “Playing Chapters”
• Formatting and layout changes
• Minor text changes (mechanics unaffected)

Here’s an example of rules leading to ambiguous interpretations: In every AP podcast of Apocalypse World that I’ve listened to, there has been confusion over how to use the Hx mechanics. As an example, here’s the stat highlighting rule:

Go around the table one last time. Every player finds the character her character knows the best, the one with the highest Hx on her own sheet (resolving ties on a whim). That other player says which of the character’s stats is most interesting to her, to highlight.

So, am I looking at the list of Hx’s on my sheet, where I have an Hx for each of the other characters and finding the one with the highest rating? Or am I looking at everyone else’s sheet where they have an Hx with me listed, and finding which of their sheets has the highest rating for me? It’s not easy to figure that out: it can be easy to lose track of who the pronouns are referring to, the concept of nonreciprocal relationships isn’t always intuitive for people, and it is easy to forget which direction the Hx “____ knows ____” relationship flows in. Is the Hx on my sheet how well I know you, or how well you know me? Usually, it’s pretty clear to the participants that they could make a good case for either interpretation so they look it up in the book or defer to someone with greater rules expertise so the ambiguity doesn’t lead to a breakdown, but it frequently results in confusion at the table.

A more pernicious situation can arise if people start using incompatible interpretations of something but don’t notice that their fellow gamer is operating an alternate (possibly mistaken) interpretation. I think I noticed an element of this in a recent episode of the Virtual Play podcast (for the specific details, listen to the episode then read my comment on the website). One of the things that many RPGs expect is for players to interact with the system in a slightly oblique way, such as by doing or saying things with your character with the expectation that another player will map that fictional action to some sort of mechanic. Even if the “sender” and “receiver” aren’t on the same page, it’s frequently possible for the receiver to think he has successfully translated sender’s input into the mechanics because the mapping functions we use in a lot of our subsystems tend to be very tolerant of unusual input: My perfectly normal attempt to indicate I’m doing X in situation A can be, with a sufficiently generous spirit, be interpreted by you as Y that kind of makes sense for situation B. You don’t notice that I’m not operation in the context of situation B since my input seems to function in your context, and I don’t notice that you’re not operating in situation A since you are indicating that you are successfully interpreting what I’m doing.

I noticed another potential source of ambiguity while listening to the Roo Sack Gamers use the Fight! mechanic in Burning Wheel Gold (a system with a reputation for complexity). Since our hobby involves a lot of talking, we tend to adopts shorthands when speaking, i.e. we drop “unimportant” details since we expect people to pick them up from context. It’s not unusual for someone express a thought like “I’m not sure if weapon lengths matter on this roll, but I know that when weapon lengths apply I get 2 bonus dice for using a sword” as something like “I get +2 on this, right?”. If the person asking the question and the person answering it are thinking in the same context that works fine. However, if lots of elements of the system can translate into a +2, it’s easy for the person answering to assume that an entirely different question has been asked (maybe you’re getting a +2 for a successful maneuver on a previous action), and they might say “yes” to the question they think the uncertain player is asking. Hearing affirmation will entrench the idea in the questioner’s mind, even though it was never fully articulated and might have been rejected if it had been. If the players aren’t on the same page they can drift further and further apart until they finally notice that something is breaking down. Once they do, they’ll need to look backward through multiple steps to find the source of the problem, which makes this seem like a “complex” situation to resolve.

I’m not sure I have any compelling conclusions to take away from this line of thought yet (other than to bring some new skepticism to the idea of “universal” mechanics within a game, since mismatched contexts will be easier to detect if different mechanics require different vocabularies), but I wanted to get some of these thoughts down in a more concrete form.

Parallel real-world player tasks are a human interface design issue

I’m playing a Fighter, and in both sessions my Strength has been highlighted (which is probably what you’d expect). The most common way for me to roll my strength is when I Hack and Slash. Since I have a high strength, that means I’m almost always going to get the 7-9 or 10-12 result, which means I need to roll damage. But whenever I roll my highlighted strength I need to mark experience. I use the same hand to roll dice and put marks on my character sheet, so I can’t do those things simultaneously. Rolling damage is the more important thing (since another person is waiting on that result) so I usually do that first. As a result, I often forget to mark experience because I’m paying attention to the damage and the accompanying description and getting caught up in the momentum of what’s happening. In later lulls, I’d often find myself wondering whether I had remembered to mark experience and having no way of figuring it out, leaving me with a vague sense of being cheated out of something I deserved. I think having a game process that asks a player to do two independent things in parallel is asking for trouble in a game’s design (I ran into a similar problem when playtesting my own game and felt the issue was important enough to completely revamp major subsystems). As I understand it, Apocalypse World usually puts the job of interpreting the results of a move into the GM’s hands after the 2d6 hit the table which leaves the player less mentally encumbered for remembering to mark experience.

Am I indestructible?

My Fighter has Armor 2. Most of the foes I faced in last night’s game were doing 2 points of damage when they were hitting me. Let’s do some math: 2 – 2 = 0. It became immediately obvious that they weren’t really threats to me, at least in terms of hit points. This caused a bit of a fictional disconnect for me when we got near a goblin army camp: on one hand an overwhelming horde of goblins ought to give my character pause, but on the other hand I know (or at least think I know) that nothing they can do to me is actually dangerous. I’m skeptical that flat monster damage and flat armor ratings are doing good things for the design. I wonder if there should be something more like AW’s Harm moves, maybe something like: “When a blow strikes you in a place you are armored, roll +Armor. On a 10-12, take 2 less damage. On a 7-9, take 1 less damage.” Or maybe take an idea from Iron Heroes and make the damage reduction from armor into a die roll: bigger dice for better armor, but you still might get a 1. This might just be a symptom of a different problem, though, which is that our GM was having a hard time wrapping his head around what to do with moves besides “the monster does damage”. Maybe things would have been more interesting if he was offering hard choices, etc., on a miss.

The basic miss/hit/strong hit mechanic is a lot of fun and works well to support the dungeon-delving situation. When were were using the mechanics well things seemed to have a good, fun flow. I felt like there was a lot of room for creative descriptions and vivid imagery, but also a strong mechanical foundation that was well aligned with what we were trying to do.

I like the “describe what you’re doing, and if it matches a move we’ll roll those dice” mechanic that AW and DW use, but some of the moves are more natural than others. Spout Lore seemed the clunkiest to me. There didn’t seem to be a way to naturally work in the fiction that would translate into Spouting Lore, it felt a lot more like I was pressing the Spout Lore button and having a power go off (I was strongly focused on this because my INT was highlighted as a non-wizard character and this move was my only way to get experience for that stat). The most natural way I was able to do it was to say a lot of “I heard about this when I was in …”, but that still felt pretty awkward. I find it hard to telegraph that I’m angling for this move because the “fiction” that the move corresponds to is so abstract. It’s easy come up with things to do that will obviously translate into Hack and Slash, but “probing my memory” is hard to make into something that would be naturally foregrounded with my normal storytelling instincts.

The fighter’s Bend Bars, Lift Gates move on the other hand was a lot of fun. In the first room we entered in the dungeon there was ancient magic that was trying to keep us from using our weapons, but my fighter wasn’t going to tolerate that kind of nonsense. After some initial frustrations with more mundane approaches, I was able to turn to Bend Bars, Lift Gates as my “all right, it’s time to stop screwing around!” move and start smashing stuff and eventually convinced the magic to back off. The scene turned fun and the situation helped my character concept to gel — when I first started out he was mostly just a seen-it-all mercenary guy without a personality. Now I know he’s really ornery and won’t give in to anyone or anything.

The jump in hitpoints going from first to second level was pretty dramatic. When we first started playing my fighter only had 8 hitpoints and it felt gritty and dangerous, especially since I had been taken down to only 4 after a few hits. But leveling up gave me 7 more, and it felt a little incongruous. The fictional situation hadn’t really changed, but suddenly the world was half as dangerous as it used to be. I know this is kind of a D&D thing, but I’ve never played D&D so I don’t have any nostalgia working against the feeling that the psychological signals this was sending didn’t make any sense.

The idea that I could use the Multiclass Dabbler advancement to pick up the Wizard’s Spellbook, Prepare Spells, and Cast a Spell moves seemed a little strange to me. It feels like I’d be taking the Wizard’s entire shtick if I did that. Maybe there’s some subtlety I’m missing, but that feels like a bigger deal than dabbling to me, and like a bigger deal than some of the other available advancements. I suppose they’re not all meant to be equal, but the difference in degree felt a bit odd to me.

In some ways the Improved Weapon advancement is a little odd, too, since many of the enhancements seem like they would be hard to see developing “in play” on an inanimate object. I also made some of my chargen-time enhancement decisions based on aesthetic decisions, so I’m not inclined to change them during play — I decided that I didn’t like the image of a sword with hooks and spikes so I didn’t pick that one (even though the mechanical effect was appealing), but for as long as I have the Improved Weapon advancement available to me it’s going to feel like I need to constantly revisit that decision. In theory I like the idea of being able to pick an extra enhancement when I level up, but the particular choices I’m looking at are making this feel less cool in practice than I think it ought to.

I’m not really feeling the “mark XP when you use highlighted skills” mechanic. I know there’s been some commentary on the net about other people not liking this subsystem either, and I’ll probably need to check out some of those discussions. I think part of it is that focusing on different stats in Apocalypse World can lead the drama in different and interesting directions, but the dungeon exploring premise of Dungeon World already focuses play quite a bit, so the highlighted stats seem more arbitrarily mechanical without a lot of story payoff — I’m still a badass fighter exploring a dungeon and chopping heads off goblins: highlighting my INT will cause me to Spout Lore more often but it probably won’t have a meaningful impact on the actual events of play. At least that’s how I felt in the first session.

I’m really excited to be playing the game and am really looking forward to the next session next week.