Game design theory rambling, part 2

The other thing that I found helpful in examining Dungeon World is the combat system, which almost doesn’t exist. That’s the part of my A Wilder World designs that most reliably falls apart, where the tension between two very different design/gameplay philosophies I am pretty passionate about reaches its breaking point.

At some point in recent months, I saw a Twitter thread where was Mike Mearls spoke about how the design paradigm for D&D shifted as things like podcasting and streaming of live games became a thing. Before that, the internet age experience of roleplaying games was based on the culture of gaming forums, which delve into crunchy minutiae and care about powergaming and balance in a very mechanistic, programmatic, nitty-gritty way that doesn’t really match up to the experience of most actual play in the wild.

With that as both the “internet focus group” for the developers and the major way for new players to encounter games, game development alternately pandered to and tried to predict and contain the theorycrafters, and the game (in my opinion) suffered for it.

But with the advent of life play streamed through people’s screens and speakers, the more open, imaginative, character/story-centric style of play that is far more typical “in the wild” became accessible to anyone. The new standard model that most new players experience is less “*snort* You put points in Charisma as a Dwarf Fighter? Your seed is weak and you will not survive the winter.” and more “Can I light the chicken on fire and shoot it out of my crossbow into the bucket of exploding sardines?”/”I don’t know, but it sounds like a good idea. Roll and find out!”

One of the key tweets in the sequence said:

This is HUGE because it shifts the design convo away from “How do we design for forum discussions?” to “How do we design for play?’

(The tweets aren’t threaded, sadly, but you can find them with this search, if you scroll down to the numbered tweets.)

My initial designs for A Wilder World were based on the assumption that I needed deep, hard mechanics and then people would adjust it for actual play at the table if they wanted to. After getting into 5th Edition, and realizing that every edition of D&D I’d played, we played it certain ways (taking the mechanical translation of spells and abilities as representative examples of the sorts of things they can do, for instance), I realized that I was more interested in designing a game that was written the way I played.

But I’ve been working with the “descriptive rather than prescriptive” approach to qualities for well over a year now, but I hadn’t ever really made peace with it until I read Mike’s thoughts and realized why 5E had triggered this decision.

And even after that, I’ve had the hardest time figuring out what to do about the combat system. The “prescriptive” draft of A Wilder World had quite a detailed one, with all sorts of damage types and conditions and maneuvers. The descriptive drafts have always tended to drift more like that when I turned to combat. I’ve always taken it as a matter of faith that combat rules have to be more complicated and nitty-gritty than everything else, in a game that models action-style combat, even as I loathe the trope of a CRPG-style freeze-and-dissolve as a tactical mini-game replaces the actual roleplaying in process.

Dungeon World comes as close to subverting that as any game I’ve seen that’s clearly modeled around emulating D&D ever has, and in fact as close as any game I’ve seen with a formal combat system at all. Dungeon World has only a small one, but it has one. It’s so unlike anything I’ve ever seen that it was really hard to get a feel for the rhythm of it, how it would play out, reading the SRD (which was why I spent so much time looking for good examples of actual play).

As small as Dungeon World’s combat system is, I do think the game itself is a bit overly formalized, with its focus on formal “moves” as the building blocks of gameplay/narrative, and character classes that progress in a special-move-based approximation of D&D character tropes. As I said in my previous post, it’s almost the kind of thing I’m looking for.

But its approach to combat, combined with the revelations from the aforementioned tweets, basically became permission to just throw out the combat minigame. The basic rhythm of the game outside of combat is: situation arises, players decide what to do about it, their own rolls, common sense, and narrative negotiation determine the outcome. Why must combat be any different? If they plan well and roll well, they escape danger. Bad rolls can result in damage, danger, or cost.

“I want to use my whip to trip up the bandit captain.” is the same type of plan/gambit that “I want to use my whip to swing across the gap.” is. If it makes sense the character could pull this off, roll for it and resolve appropriately. There’s no actual need for separate rules for hit rolls, initiative, turns, etc. If the roll goes poorly, the consequences will arise from the situation; in a combat situation, this would include the possibility of taking some hits.

And that’s really the key: treating combat as a situation rather than a system or mini-game.

The part of my brain that’s on the “forum play” mindset still resists the idea that things can be that free-form. Surely if the players don’t take advantage of it, a bad game-runner will and wipe them out. Surely if there aren’t rules built to handle things like lashing an enemy with a whip and pulling them off their feet, players won’t think to do it and the game-runner won’t know how to resolve it if they do. Surely, surely, surely…

But that part of my brain no longer registers as definitive. It’s also made me a bit freer about saying yes in games of actual D&D where, as much as I follow the “representative examples” school of thought, I am more fond of limits than my players often are, particularly when it comes to things like spells and supernatural abilities.

Let’s do some game design theory.

An attempt at a breath of normalcy in abnormal times.

I’ve been picking apart Dungeon World lately, a game that has been recommended to me many times (mostly by snide D&D players who, hearing that I care about character and story, cry, “Why don’t you just go play Dungeon World, then?”) but which I’ve never played. The SRD for the game was not a very intuitive introduction, though, so it took me finding some podcasts of actual play to get a handle on it. I think it’s a solid game

Like Fate (another frequent recommendation), it does approach to some of my goals without actually quite fulfilling them. I do enjoy the dice mechanics, and the way they’re integrated with the narrative negotiation… a concept my mind initially rebelled against at the level in which it’s present in Dungeon World, but in which I’ve come to be a believer.

So for my latest iteration of A Wilder World, I started with a dice mechanic very similar to Dungeon World, then walked it back to be closer to the last version of AWW’s. Here’s where it stands right now.

A check is made using 3d6. For a very simple, low complexity, low stakes pass/fail, you can treat a 9 or lower as a failure and a 10 or higher as a success. This gives a 62.5% chance of success; anec-datally, around 60-66% success/reward seems to be the tipping point where “random” things are more fun than frustrating.

Any modifier to difficulty is applied to the roll itself (as in the optional rule in Dungeon World). If you have an attribute to apply to the check, it is also a straight mathematical modifier. The scale is: 0 is broadly average, any deviation is meant to be notable, a typical heroic PCs have positive scores of up to +3 at the start of their career.

The game uses an advantage mechanic (similar to the concept in D&D 5E) to represent both significant situational modifier and the main effect of Heroic Qualities. If you have a quality like “Acrobat”, you have advantage on any check an acrobat would have advantage; the current version of the game, being light on mechanics and heavy on narrative negotiation eschews detailed nuts and bolts for qualities and instead offers broad-strokes descriptions of the sorts of things they might cover, with the idea that the player and the Storyweaver will work out exactly what it means.

The basic rule is “If you would expect that a character possessing this quality in the sort of story you’re telling might be able to do it, you might be able to do it.” And if it’s something that anyone could try but your quality would make you better at, you have advantage when doing it.

The exact effect of advantage is to add one die to the pool for the roll, with you still taking the top 3. Advantage shifts the odds without shifting the spread, in other words. And while you can stack multiple advantages, there are diminishing returns.

Benefits of this system over Dungeon World: the dice are less swingy with another one in the mix, but there’s a wider range of possible results, and outright failure stays on the table longer even as it becomes less and less likely.

The simple pass/fail check can be given more gradation, either informally by simply saying “higher success is better, lower failure is worse” or using a “color table” where 9 or lower is red, 10 to 12 is orange, 13 to 15 is yellow, and 16 on up is green. If you’ve seen Dungeon World’s success table, the idea behind the colors is similar: red means outright failure or very costly success, orange means tenuous/partial success, or success with a complication or cost, yellow means simple success or the possibility of greater success with a cost/complication, and green means unqualified success with some unexpected benefit.

The whole thing is very interpretive, with the idea being that in situations where they apply, the Storyweaver might offer the player who makes the roll a choice between the two interpretations for the color they rolled (except for green, which has a single meaning) and possibly soliciting a suggestion for what the cost/complication/boon is, if applicable. Tables/groups that prefer a more straightforward game where the game-runner controls the narrative of the world (apart from player character actions) can run it that way, with the Storyweaver making the decisions.

There are some <BLATANT LIE>fascinating</BLATANT LIE> statistical minutiae regarding the exact odds of different color results and how they shake out when you have various combinations of attributes and advantage, but I won’t spell them all out here.


Monday Monster: Wasp Mount

In Saturday’s game, I threw a group of insectoid bandits at the party. The bandits were a large number of thri-kreen, bulked up a level with the quick-and-dirty CR-increasing method of increasing their attack bonuses and defenses with +2, plus giving them a better multiattack routine. Two of them had mounts that I adapted off the cuff from the existing giant wasp creature in a similar fashion.

Oh, I’m realizing as I type this that I based the experience reward for the encounter off the listed CRs and not the adjusted ones. Whoops! That’s a risk of such on-the-fly adjustments. I thought the number I wound up with sounded low for the challenge of the fight. I’ll have to make that right at the start of next session.

Anyway. I wasn’t fully satisfied with my wasp adjustment, because the existing giant wasp is a medium creature and one large enough to carry humanoid riders would differ by more than few points’ hit/miss chance. So this afternoon, I’ve taken it upon myself to make a more complete wasp mount write-up. As always, the stat block was created using critterDB.

I started with the existing giant wasp and scaled it up a size, increasing its hit dice size in accordance with the guidelines for such and then giving it a few more. I also increased its Strength and the damage dice of its attacks, to reflect is larger size, then upped its Dexterity and proficiency bonus simply because it’s meant to be a more menacing creature than the baseline giant wasp.

The added traits of Flyby and Lifting Capacity are simply there to make things interesting. One way to have a good mount is to have a creature that is not particularly dangerous or even that compelling on its own, but which when ridden by another creature almost exponentially enhances its deadliness. A flying mount that can avoid opportunity attacks hits that niche, even before you add a wicked poisonous sting.

The “lifting capacity” trait is mainly there for story/non-combat purposes (giant flying wasp abductions!) rather than in-combat uses. As appealing as many people seem to find “drop them from great heights” as a combat tactic, round-by-round it’s almost certainly going to be better to have the wasp and rider make flyby attacks, particularly if the rider has multiattack.


Play #RealmLike in your browser!

So, I didn’t mean to spend the whole day bashing around with #RealmLike, but enough potential players told me they wanted in but didn’t want to have to mess around with downloading an external program that I moved making it web-accessible up a few notches on the priority table. I was due a free web domain from my host, so I snagged and threw it up.

A couple of caveats: the “off-the-shelf” web interface ignores a lot of my in-game text formatting, so a lot of things look clunky and weird (and not in the deliberate style of the game). It also includes a lot of extraneous bits (like sound controls). The next iteration of the browser interface will not have those, but I’m running into a few tricky issues with building it. The [T] to Talk command does not function in the browser; you have to manually click the cursor into the chat box. I really want to fix that because it is my intention that this game be 100% playable with keyboard.

So the game is in early development, and the browser interface is in even earlier development. It is, however, playable. Upsides: making the game compatible with it also involved streamlining the training menus, and since I was redoing them anyway, I went ahead and added some things I’d wanted to. Now when you choose something, it gives you a description of what you’ve selected and asks for confirmation, and loops back to the main training menu if you still have selections to make.

Basically, every day the game gets a little more intuitive in response to player feedback.

You can now play #RealmLike on your own computer!

Caveat: single player only so far. The dungeon dimensions and other parameters that will eventually be customizable from within the game are currently hard coded in; during this early phase of testing, I’d like to limit the amount of x-factors involved when I can’t see what’s happening, in order to keep the testing useful.

Downloading the game requires a BYOND account (still free!) and a $5 purchase of RealmLike through BYOND. It’s called a subscription because that’s what the platform uses as its model, but it’s lifetime. Think of it as similar to buying a game on an Early Access model. Your one-time purchase nets you any and all future updates, and you will get notifications through BYOND when a new version is uploaded.

As an added cherry, subscribers get more character slots on the public test server and may occasionally have access to a separate private test server where newer things are being tested out. Future character classes and races will likely see the light of day there first.

The stable build that’s up for download now includes all the innovations added as a result of today and yesterday’s testing, including:

  • Longer period before scarring sets in from vicious wounds and burns.
  • Consumable items to remove vicious wounds, burns, and scars (bandages, aloe salve, and miracle cream).
  • Consumable items for navigating the dungeon.
  • Some basic color signposting and random cosmetic dungeon features to help you navigate from a level entrance to a level exit.

There is also a vastly improved communication/social system, though it’s a bit superfluous in the offline only version of the game.

#RealmLike: Minor updates

Well, a few intrepid souls have logged on and tried RealmLike out during the day. These folks will have made things a little easier for everyone who tries it out tonight, as through their feedback and my own observations I have identified a few troublesome areas that didn’t give me any problems but which weren’t exactly intuitive to outside eyes.

Stairs were previously labeled “v stairs v” and “^ stairs ^”, to indicate directions. Since the first stairs you encounter are pointing downward, this means the first thing you see is stairs surrounded by a letter v. Two of the three people I witnessed playing the game did the logical thing and pressed V to use the stairs. This is doubly confusing, as the Climb command (Z) is the only one whose key shortcut is not directly derived from its name. I’ve now relabeled the stairs as “Stairs [Up/Down] Press Z”.

The trainer in town is currently separate from the training menu you can bring up by pressing +. This is because the trainer can teach you new classes, but as this was the only intended function of the trainer when I implemented it, talking to one while you’re not ready to gain a level gave a message saying to come back when you were ready to train, even if you had other training to do. I’ve now adjusted the trainer to bring up the normal training interface when you’re not leveling. Long-term, I’ll merge the two interfaces and just give you the extra options when you access it through a trainer.

These two things created a potentially frustrating ping-pong experience for players, as you could spend precious new game energy time trying to figure out the arcane secrets of the stairs only to be told to finish your training, and then go talk to the trainer and be told you’re not ready to train. I’ve now ironed out those wrinkles.

I also made secret passages show up more distinctly; the subtle color shift when your perception is high enough to notice them wasn’t really very noticeable, and since there’s an internal “perception check” being made, a real-life one isn’t called for.

A couple of more vicious monsters had their levels labeled incorrectly so were spawning too frequently on the first level. This has been fixed.

Finally, I made a sort of “initiative roll”/delay when a monster moves adjacent to you so that it doesn’t immediately attack the same tick if its attack is cooled down. There’s no visible wind-up, but if you have auto defend on (it is on by default), your character may score the first strike, depending on their attack speed and reflexes.


So you all know I dabble in game design at an almost constant background level.

Computer. Tabletop. Whatever.

I do it most often when I’m overwhelmed by anxiety or can’t sleep in the middle of the night. It’s like meditating on an orderly world, and it somehow uses fewer processing cycles at its most intricate than creative writing or almost anything else does.

So I started a game design experiment the night before flying out to WorldCon, a bit over a month ago. And then I ignored it, only to pick it back up and poke at it during the last few of the long nights we were visiting the hospital. And then during the week after that or so that I was bombed out of my mind on cough syrup and allergy meds, I really threw myself into it.

And so now I have created RealmLike, a retro-dungeon crawler and minimally multiplayer online RPG. The name “RealmLike” derives from RLML, for “Rogue-Like/MUD-Like”, naming its two most immediate influences. It’s a functional game engine that is capable of generating a dungeon 250-some tiles across and 255 levels deep in a matter of seconds (though it’s currently set to a slightly more reasonable size, five levels deep), which can then be explored in classic dungeon-crawl style with some modern sensibilities (like skill trees, multiclass, and hotkeys for spells and abilities).

This kind of game is the sort of thing I’m always tinkering with. My projects usually stall out or fall apart because even working in a retro palette with low resolution pixel graphics, my needs fast outstrip my abilities. RealmLike gets around this by fully embracing the abstraction embodied by its oldest progenitors, though with a 21st century twist enabled by higher screen resolution:


21st century text-based graphics don’t have to stick with “d” for “dog” and “g” for goblin. We have enough space to spell things out.

Game has been coming along well, I’d like to bring it along further, but it could really use more hands and eyeballs on it to make sure everything in it works as well as I think it does. If you’re interested in checking it out, I am going to be leaving a server running.

This game is made with the BYOND programming suite and currently requires BYOND’s client program to connect to it and play. You can download and install the suite here (it’s free!), after which you will be able to join the game from its hub page.

Controls are entirely keyboard based: arrow keys to move, keyboard keys are single-letter commands such as U for Use or G for Get. There’s a help page that is displayed when you first login (F1 can bring it up again) that lists all of them. Character options are a bit limited right now; four character class and four fantasy races (if you automatically and immediately know what those four are, you are probably right in the target audience) and they are capped at level 5 for the non-spellcasters and 3 for the spellcasters. But there’s some variety within them, with a minimum of two skill sets/sub-classes for each and some additional customization options for Clerics.

The game is online, so if you join the server, you may theoretically be playing with other people. You can even form a party with ‘P’ and share experience (automatically) and other resources (voluntarily). Feel free to drag your friends along. There is no PvP right now, and while permadeath is an option, it’s currently turned off.

I plan on including a full level progression, more character classes, more environments (and more details for the existing one), magic items, crafting, all the wells and bhistles, save for graphics more sophisticated than stick-on labels. But one woman can’t playtest a game herself, particularly a multiplayer one. Think of this as open alpha testing, like a Steam early access game.

You can leave feedback on the game’s hub forum, or by using the GM Chat (hit F10). I’ll be “AFK” from the game most of the day but messages sent there are logged.

Monday Monster: Flame Bear

My current main D&D campaign has a theme of nature and magic running amok, so I’m always on the lookout for twisted versions of natural creatures. The flame bear is an example of such a critter, along the same lines as the somewhat iconic winter wolf (which I’ve already used): smarter, bigger, elementally-infused version of a famliar animal.

Flame Bear

Note that I’m not too worried about my players seeing these stats, as most of the monsters I use are commonly available. Also, my players tend to prefer fewer battles with more interesting foes, so I’m working on a legendary version to throw at them.

8 Hour Game Hours 7 & 8: Not quite the home stretch.

Okay. So. I very broadly hit the goal of creating a rules lite, diceless game system that could technically be played, but I’m going to give the final results a kind of mixed rating? The equipment system is not there, there are references to special abilities but no framework for such, there’s no elements of narrative control that I was hoping to include. What I have is less an elegant foundation on which an end user could build whatever they wanted and more a stripped-down chassis which a skilled mechanic with the right tools who is willing to get their hands dirty could use to build something, with a little effort.

Still! It’s something. And I think the timed/deadline approach had a positive impact on my design process and keeping me focused. So, I’m going to extend it. In a previous update I was talking about extending the system through a series of similar focused work days. I’m going to give the base system itself a little more time, though. Specifically, I want to see what I can come up with in 24 hours. Not all at once! I have too much going on for that. I doubt I could find two more 8 hour work days in the next few weeks.

But a few half-days, here and there? Totally doable. So I’m going to log the time I spend working on it and see what I can come up with in 24 hours total. This 8 hour session gave me a more realistic idea of what I can do in 8 hours, so I think a (distributed) 24 hour design schedule is totally doable.

Here’s the final updated link for now. Again, this is with no polishing, minimal editing for order or clarity, and not at all the preferred newbie friendly, strictly-defining finished version. Future updates will still be tagged “NDO” but not “8 hour game design”, because, not 8 hours.

8 Hour Game Hours 5 & 6: Skills and Such

Okay, so, my attempt to update my progress on the hour has turned into every other hour. Skills took me longer than I expected, in part because I indulged in my love of specificity and in part because I had to stop and consider some sub-systems for performance skills (I loathe the idea of a general all purpose entertainer skills) and for knowledge skills that are culturally bounded. I’m pleased with what I came up with for both cases, although as with the entire document as a whole, the presentation will probably need to be unified and cleaned up.

I have two hours left in the day. I suspect that to reach my goal of having a completely playable, self-contained if simple game, I’m going to have to make the equipment system more bare-bones and abstract than I would like; i.e., define weapons and armor in terms of general classes, and pretty much leave everything else up to improvisation. This is partly because of time constraints, but partly because I’m trying to make a base system that is universal and agnostic to setting, which means either having an incredibly large and detailed equipment sytem with everything from stone knives to phaser rifles (not going to happen in one afternoon) or leaving the equipment system skeletal, to be filled in for particular settings.

With that in mind, I’m going to proceed and try to wrap this up in the allotted time as planned, but I’m setting further goals for myself: take time to develop this bare-bones system into something more fleshed out for particular genres/settings, using a similar approach to this. I’m going to start by expanding my 8 hour game design into 24 hours  (not concurrently! Rather, three working days) to turn it into a high fantasy adventure game, with appropriate equipment system, magic system, etc.

Updated link.