Monday, December 9, 2019

Roguelike Advice for Tabletop Games from @Play and Golden Krone Hotel

John Harris of the @Play blog and @Play column writes about "rougelike" videogames. Since I am somewhat interested in procedural generation in tabletop gaming, there are a few of his columns that I particularly like. There are also a couple contrarian pieces from Jeremiah Reed of Golden Krone Hotel, and some ASCII art I like from Uncaring Cosmos and Imminent Demon Engine, and some links at the end for resources for making ASCII and pixel art.

image by Uncaring Cosmos

image by Uncaring Cosmos

Purposes for Randomness in Game Design is about reasons to use procedural generation instead of "set" content in a videogame.

- to make multiple playthroughs of the same game interesting
- to offer a game some resistance against "spoilers"
- to challenge players' skills by asking them to deduce things about the gameworld
- to create emergent narratives that wouldn't arise any other way
- and to create emergent complexity by randomly combining basic elements

In tabletop gaming, I would add the reason that it allows the gamemaster to discover the world at the same time as the players. I would also add that one of the challenges of procedural generation at the tabletop is that proc gen makes it harder to offer players a meaningful, clue-filled environment where they can successfully deduce what's around each corner - so it's quite interesting to me that he lists that as a strength.

Eight Rules of Roguelike Design is kind of a manifesto for rougelike gaming. Most of these seem like good advice for any dungeon, and a few at the end are especially relevant for resource-management, exploration-style gaming. It's worth remembering that if you want your players to interact with the mysteries of the dungeon, they'll be more inclined to do so if those mysteries aren't usually harmful, and if even the harmful ones aren't instantly lethal.

Some of the advice about unidentified items initially struck me as being kind of narrow and genre-specific, until I remembered that item identification is a kind of mini-game inside Numenera and Mutant Crawl Classics, among others.

- no player character should be immediately killed by a single monster attack
- no player character should be immediately killed by testing an unidentified item
- magic items should require testing to identify, even for players with a lot of system knowledge
- each magic item type should have enough potential effects that testing it during combat is potentially beneficial but also potentially harmful
- magic items should have both benefits and penalties (or at least limitations) so that they present interesting choices
- because magic items have both upsides and downsides, no item should ever be completely useless
- exploring the dungeon should use up a resource so that players aren't able to explore indefinitely
- as you explore deeper into the dungeon, monsters should become more dangerous a little faster than player characters become stronger (so that magic items become more important over time)

Towards Building a Better Dungeon is all about the things tabletop games still do better than computer games. There are a number of experiences and mechanics that I've noticed work better for single players than they do for groups, or that work better when a computer is handling the numbers than when humans are, so it's nice to see someone from the other side praising what works better in our world.

It's also interesting to see which aspects of of D&D he admires. It's many of the same things you see praised on OSR blogs, for example. Although the staircase thing seems like it's an artifact of the way rougelike games randomly generate their maps - it seems so common-sensical to me that I struggled to even write the one sentence summary, but apparently it's an issue for them. There are other elements of old-school D&D that would be difficult to replicate, such as factions of monsters that want to recruit you into their internecine conflicts, but what he focuses on are mostly the elements that would enrich solo play.

- D&D has varied, interesting that are placed deliberately rather than randomly
- monsters in D&D come in different sizes, from small to large
- old-school D&D requires narrative searching to find secret doors
- on multi-level D&D maps, staircases are placed consistently in relation to one another
- despite its difficulty magic item identification is actually easier than in Gygaxian D&D (I suspect roguelike games also don't contain Gygax's, uh, rogue's gallery of look-alike monsters that exist solely to punish his players for adopting the very same playstyle he pushed on them. Also wait, someone is envious of this?!)
- you can't play roguelike games with your friends the way you can with D&D

Meanwhile over at Golden Krone Hotel, we get Things I Hate Sbout Rougelikes: Bog Standard Dungeons, which is, at least kind of, an argument against continuing to imitate D&D and Lord of the Rings in new games. My reading of this isn't that he's criticizing vanilla fantasy per say, but rather, that he's calling for more new games to employ a strong consistent theme that's not the same vanilla fantasy you see everywhere else. Of course, new games like Torchbearer, Dungeon World, and Forbidden Lands all developed large followings by selling "vanilla fantasy but with different rules" - so what's good artistic advice and what's sound marketing strategy might differ here.

There are three parts to his complaint:
- high fantasy is vanilla, and more importantly, it's overdone
- kitchen-sink bestiaries end up full of monsters that feel inappropriate or out of place
- a few "goofy" elements will quickly make an entire setting feel goofy (Which might be an argument in favor of going full-on gonzo. One joke monster just spoils the mood, dozens of joke monsters actually become the mood.)

I actually kept thinking about Jack Guignol's In Defense of Vanilla Fantasy while I was reading this. Because they initially seem like they're going to be in disagreement, but in some ways, I feel like they're two sides of the same argument. After all, when Jack says "they make vanilla so we don't have to", the argument here seems to be "they already HAVE made vanilla, so why do we keep making it too?" James David Nicoll has an ironic version of this plea, when he begs his readers to please, please "give the Tékumel and Gormenghast costumes a rest." Of course, Jack has a rejoinder to that, "vanilla might just be what people actually want" - like I said, there might be sound business reasons why so many game-makers keep making new vanilla games.

Even the Old School Renaissance has only one really weird megadungeon in its top five - Anomalous Subsurface Environment. Three of the others are high fantasy - Stonehell, Dwimmermount, and Castle of the Mad Archmage - and they all start out vanilla at the top and really only end up getting strange near their final levels. Barrowmaze is built out of basically vanilla components, but it has a narrow, consistent theme, and fills up its space by offering variations on that theme rather than a funhouse of new ideas. The biggest change as you go deeper is the slow shift from undead to demons.

Settings with a lot of novelty can run into the problem that "when everything is weird, nothing is weird." But the call here isn't for random weirdness, it's for a consistent theme that's simply a different theme than vanilla high fantasy. If it feels like you have a "kitchen-sink" full of monsters, if a handful of your monsters feel inappropriately "goofy," then the problem isn't that you have too much weirdness, it's that you don't have a consistently applied theme. Real weirdness is weird precisely because it stands out against its background - whatever that background happens to be. You can still have real weirdness even in a setting where everything is (initially) strange, but it will require using only a few stand-out elements (not a sinkful) and making them at least somewhat unique, not "goofy" and not just imported from another well-known genre.

So what games does he like? Unreal World, Cogmind, Hieroglyphika, Sproggiwood, Haque, Sil, Binding of Isaac, Nuclear Throne, Spelunky, and Caves of Qud. And presumably he likes his own game, Golden Krone Hotel.
image by Imminent Demon Engine
image by Imminent Demon Engine

In Item Design: Potions and Scrolls, we return to @Play to look at good design for these single-use items. Remember, half the criteria for good rougelike gaming are based on good magic items. He argues that magic items are so important for roguelike gaming because exploring the dungeons and fighting the monsters are not, by themselves, enjoyable enough to sustain interest in the game, only the items can do that over the long term.

In both rougelikes and old-school D&D, your character is adventuring for basically the same reason you're playing the game - for enjoyment. Your character explores dungeons and fights monsters to find money and cool stuff. Money (via XP) unlocks cool level-up abilities. Money lets you buy more cool stuff. Cool abilities and cool stuff in turn let you ... uh ... explore more dungeons and fight more monsters. So these things had better be enjoyable, because enjoying using them pretty much IS the entire purpose of the game - and if your game doesn't include any level-up abilities, then the cool stuff had better be especially cool!

He feels pretty strongly that single-use items should be unidentified until they're used, and even then, only if their effect is something that the characters could notice. So if you drank a potion of monster detection for example, and there were no monsters around to detect, the potion would seem to have no effect. There are also potions and scrolls that really do have no effect, just to keep you on your toes! While apparently one of the key pleasures of solo roguelike computer gaming, I think this kind of thing probably gets tedious very quickly in a tabletop game. (Apparently the only way to get Gygax to volunteer what your magic item did was to let a Rust Monster or Disenchanter destroy it. He was happy to tell you what you just lost! Otherwise you had to go into town, hope you could find a sage with the right expertise, and then hope the sage made their skill check. Tedious!)

There are a couple elements of roguelike potions and scrolls that don't show up much in D&D, and might be interesting to try including. The first is alchemy rules that reward you for mixing potions. Unless I'm misremembering, the "potion miscibility table" in D&D basically just says, "don't mix potions, or one of these twenty bad things will ruin your day!" The second element is scrolls that let you enchant your own weapons and armor. I've never heard of someone's campaign where players routinely turn their own mundane equipment into homemade magic items. It might happen occasionally, but it sounds like a common occurrence in roguelike games.

There is one element of D&D that he points out never makes it into the roguelikes - cursed items that are look-alikes for specific magic items. In a rougelike game, you're never going to successfully identify a Bowl of Commanding Water Elementals only to try using it and discover it was actually a Bowl of Watery Death, whereas basically all of D&D's cursed items function like that.

In Objects of Collection, he lays out a whole taxonomy of items that characters can find in the dungeon:

- basic one-use item, such as food rations
- one-use unidentified magic items, such as potions and scrolls
- wearable, always-on unidentified magic items, such as rings and amulets
- multi-use unidentified magic items, such as wands
- basic equipment, such as weapons and armor
- unidentified magic equipment

He notes a few other details about each type that are interesting to me, again primarily because they're a bit different from D&D. One-use unidentified items can also include special food rations that bestow some kind of benefit in addition to fending off hunger.

Wearable unidentified items typically have a very minor effect to compensate for the fact that they're always turned on - without the computer there to remember for you, these sound tailor-made to be forgotten about during play. They can also impose an additional cost in exhaustion and food consumption. A minor increase is too finicky to consider, but I wonder if needing to eat double or triple rations would be a meaningful cost in a resource-management game?

"Basic" equipment has a random component, too. Every sword or piece of armor you can find in a rougelike game will have a secret bonus, just like the simplest magic swords in D&D, which makes deducing each item's bonus another tedious fun mini-game within roguelike play, but again, I wonder how well this would transfer to in-person play.

One thing I think is kind of neat is that unidentified magic equipment always has a predictable mundane use as well. So no matter which random magic power your magic snow boots have, they also always help walk through snow.

His final article in this series Rouge's Item ID In Too Much Yet Not Enough Detail isn't just a description of how magic item identification works in roguelike games, it's also a defense of the gameplay value of having unidentified magic items in the game to begin with. One really important thing to note, in case I haven't been clear enough about it yet, is that these unidentified items all come from a larger list, and you can find multiple copies of the same item during your game. So once you can identify an item, you don't just know "what was that thing I just used?" you also know "what will these other identical things do in the future?" The value he sees in having unidentified magic items would be considerably diminished in a game like Numenera, where theoretically every item is unique, rather than something you expect to find multiples of.

I get the sense that John Harris and other roguelike computer gamers would get along well with some portions of the OSR. He has a deep admiration for Gary Gygax and the original AD&D Dungeon Master's Guide, and he praises a number of design decisions in 5e.

So what does he thinks makes identifying unknown magic items a good part of roguelike play?
- it should be possible to use the item without identifying it first
- there should be some bad items so that using unknown items is a little risky
- sages and spells that identify an item without needing to use it should be rare
- the game needs to be difficult enough that players have to risk using unidentified items. they can't afford to wait until they achieve perfect safety to start unknown items out.
- items shouldn't be automatically identified when you use the. you only find out for sure what it is if it does something unambiguous under the present conditions.
- some item effects should be contingent on the character's status at the time of use
- bad items should have some positive use, even if it's just throwing them at monsters
- it should be possible to deduce what some items do without using them
- there should be more items in the game than can be found in one playthrough

For a contrary view, we once again return to Golden Crone Hotel for Things I Hate About Roguelikes: Identification, where Jeremiah Reed proposes a solution that's oddly reminiscent of his last one - to solve the problem by reducing its complexity. Previously, he argued we could "fix" funhouse dungeons by applying a theme to limit what kinds of monsters can appear. Here, he suggests that we can fix magic item identification by reducing the number of possible types that any particular unidentified item could be. I'll come back to that in a second.

His critique of roguelike identification is probably not that hard to guess, but let's look at it briefly anyway. He starts with a series of examples showing the many ways a player can die while using an unidentified magic item, either because the item was directly harmful, or because it provided no help in a dangerous situation.

- Outcomes like that are especially punishing on novice players. Experienced players should be rewarded for their accumulated system-knowledge, but it shouldn't be impossible for someone without that knowledge to play the game.

- It encourages item-hoarding (more on THIS in a second, too) which both makes the game more boring and makes it harder to survive.

- It makes using unidentified items feel like a trap, even though it's not supposed to be. (There's a similar problem in negadungeons, although there it feels like everything's deadly because truly everything IS deadly and will kill you if you interact with it.)

- And for all that, there are enough meta-game tricks that sufficiently system-knowledgeable players can accurately guess what most items are with in-game identifying them. Which seemingly defeats the purpose of making them unidentified in the first place.

So as a solution, he proposes that unidentified potions come in groups of three - each potion is recognizable enough that it could be one of three different things. As an example, he shows a character considering drinking a potion that might be a ration of honey, an antidote to poison, or teleportation in a bottle. The idea here is to encourage players to take more risks with their characters by limiting the scope of their choices. You still don't know exactly which effect you'll get, but it won't just be a dice roll on a d100 table - it'll be one of three things, and importantly, you'll know the worst thing that could happen when you make your choice.

I genuinely like this idea, and I feel like it could have other applications. You see a monster at the end of the hallway. It's a skeleton, and your cleric is certain its one of three possible undead creatures. Or you find a scroll in an unknown language. Even before you translate, your wizard thinks it could have one of three possible effects. Or you enter a room know that you've just stepped on a pressure plate. Before you lift your foot, your rogue tells you the three possible traps you might just have triggered. I particularly like the thought of applying this approach to Zonal anomalies.

There are only two difficulties with putting this idea into action in D&D. The first is that it would take a bit of preparation to add in this extra potential information into an adventure that didn't already include it. The second difficulty is that without a computer to do the hard work for you, it would really take some preparation to re-randomize these associations after each playthrough. Having DM aids that are essentially worksheets you fill out in advance (like the ones Signs in the Wilderness makes) would certainly help.

Finally, all this talk about single-use items got me thinking about Razbuten's video Consumable Items (And Why I Barely Use Them). After all this talk about identifying items, it's worth thinking about what makes you want to use them. The "barely use them" problem is definitely me walking around with a full complement of missiles that I never fire in Super Metroid, or accumulating dozens of Mushrooms and Tanuki Leaves in Mario 3. Razbuten divides single-use items into two categories - "reactive" items that restore hit points or eliminate status injuries (like poison or blindness), and "active" items that proactively affect the world. He argues that most players will use "reactive" items whenever they need to, but end up saving (and forgetting!) their "active" items.

One reason he thinks this happens is that players can often pretty easily win fights and beat the game without using any items. He notes that he uses more items on harder difficulty settings, where the extra boost is the only way he's able to win fights that he can simply hack and slash through at normal difficulty. I think this goes to @Play's earlier point that roguelike games ought to get harder faster than the hero character gets stronger, which will make equipment more important over the course of the game. Having a few monsters that are much stronger than the others can encourage you to use your items to win those fights in particular - although possibly at a cost of wanting to "save up" for those fights.

Making ALL equipment temporary might also encourage players to use single-use items more freely. Instead of using a permanent item to preserve your single-use items, you might be tempted to use up a single-use item to prolong the lifespan of some of your other equipment. Legend of Zelda: Breath of the Wild does this, as does the old SNES game Brandish. This could be a little hard to track in a game where a computer isn't counting your sword-strokes, but of course you can make the attack roll do the work for you. If even the best items break on a natural 1 (and less durable items break on a wider range) then nothing is permanent, and you need to keep finding new weapons and new armor throughout the game. (That one might be a hard sell for your players though. A bronze age or stone age setting could make it more palatable.)

Making new items easy to find is another suggestion for getting people to use them instead of hoarding them. There's no reason to try to save up your items if you can be pretty sure you'll keep finding more. Perhaps you could combine that with an encumbrance system that does't LET you build up a large supply, which is more or less what Numenera does - its single use items are plentiful and most characters can only carry 3 at a time early in the game, so you have a strong incentive to use them, and little reason to save them, even though each one is unique. You might also just have to accept that most players WON'T use "special" items under "ordinary" circumstances. The key to encouraging their use, then, would be to increase the number of "extraordinary" situations where item use becomes more likely.

Having non-combat puzzles to solve can encourage experimentation, which is a point that Joseph Manola has made before. It's also consistent with my own behavior in using the slightly-harder-to-replenish "boss power" weapons in Mega Man X. When faced with a problem that has no really obvious straightforward solution, I'm more likely to start experimenting with my equipment. Probably this is true of other players as well. Breath of the Wild includes areas that you can only reach by drinking certain potions to increase your abilities, and of course Super Metroid has its various lock-and-key puzzles where specific equipment items open up whole new areas on the map that are otherwise inaccessible. Puzzles and hard monsters, then, present a pair of difficult situations where players will "dig deep" to stay alive and overcome the challenge, and so they're both perfect times to use special items.
FINALLY finally, if looking at the ASCII and pixel art from earlier got you interested in making your own, here are a few links to free tools. When I posted about ASCII art once before, several people suggested resources to me, and I wanted to share them now. Each of these was recommended by at least one person who seemed to be in a position to know.

advASCIIdraw is a free program for drawing your own ASCII dungeon maps (and presumably anything else you'd like to draw using ASCII characters?)

Oryx Design Lab is not free, but they do sell packages of pixel-art images that you can use in your own games, including ones you plan to sell. Their prices are $25-$35 for an entire collection, and one of their collections is a rougelike tileset, which I believe is what Uncaring Cosmos used for their graphics.

Open Game Art is a repository for free, open-source, and Creative Commons pixel art. All of the art is free to download and free to use, although some artists may have licenses that only allow their art to be used in free products, while others will also allow their art to be re-used in something you're selling.

Lospec has a number of resources for making pixel art. They have a nifty list of artist-submitted color palettes, sorted by popularity, and with a number of search options. They have a free in-browser pixel art program, and a whole list of resources for for making pixel art, finding software, or locating communities of other pixel artists.

Playscii is another free program for making ASCII art. This one can make still images, animation, and can be used to make playable games.


  1. "One joke monster just spoils the mood, dozens of joke monsters actually become the mood."

    I'm getting that tattooed in reverse across my forehead.

    1. Josh, I was kind of thinking of you, Anomalous Subsurface Environment, and Operation Unfathomable when I wrote that.

      The humor in your stuff is consistent, and it's of a type that fits with the mood at the table. I don't have a general theory of humor in RPGs, but I would consider your stuff a good example of humorous elements done well.

  2. I should note that on the OSR Discord, @Chameleon argues that Nethack actually DOES have cursed items quite similar to D&D's.

    I expect that John Harris is accurate in his description of the game "Rogue", but it's possible there are details he's unaware of in some other roguelike games.

    So, thank you for that information, @Chameleon!

  3. Blimey, thanks for putting this together! Is it in service of a larger project (for example, will this be part of a series)?

    Either way, you've inspired more art from me:

    Oryx Design Lab is fantastic, btw, and I've used a lot of his art (particularly for battle sprites). The overworld map images you featured are mostly me, though (well... completely ripping off the Ultima games, of course).

    1. Uncaring Cosmos, I don't have a series planned right now, but like you said, the artistic interface between our analog imaginations and our digital tools is a fascinating topic, so I'll probably post more about it in the future.

      This already probably could have been two posts (one about the relationships between roguelikes and D&D, and another about item identification) and I have an enduring interest in procedural generation of game content, so ... yes, there will be more.

      I think you must have been the one to recommend Oryx to me, and actually, it was yours and Immanent Demon Engine's art that got me thinking about writing this. It was like I hit a critical mass of too many things I wanted to shared, and put it all together into one post.

    2. I love, love, love the old school Ultima look - you were able to do all that on Oryx as-is, or did you create the maps from scratch?

    3. @Anne - I'd love to read more from you on the topic. It's really something that interests me a lot, and I think there are probably some interesting things that can be pinched from roguelike development (and the history of RPG video games) and used in tabletop games.

      @Michael S/Chgowiz - Cheers! The overworld maps are mostly my own work; mountains, hills, plains, oceans, ships, forests, castles, towns, signposts, the avatar, etc. (though, obviously, I was imitating the Ultima style).

      Many of the town and dungeon details (like chests, stone walls, doors, etc.) come from Oryx.

      All of the battle sprites (plus battle backgrounds) come from Oryx or others on I may have "pixelfied" some sprites, or tweaked them, but that's it.

  4. This is a neat post, I need to dig through the essay's still but thanks for posting it. I generally think of TTRPGs influencing CRPGs, and am wary of influence coming the other way - simply because the mediums do such very different things.

    I certainly don't agree with all the CRPG designer tips that you brings up here - but I'm driven to read more. Thanks! I will likely have curmudgeonly, groggy thoughts on all this later.

    1. Gus, I agree that it's worth being wary of the influence of solitary computer gaming on the D&D experience ... although I suspect that a LOT of influence has been flowing in that direction for quite some time now.

      I've been kind of training myself to ask "would this work better if a computer did it?" whenever I think of a new mechanic, and frankly, whenever I'm looking over other people's suggestions for rules.

      I think it's worth looking at roguelikes for a couple reasons though. They use procedural mapping and content generation, which I find interesting to do at the table. They also emphasize equipment and minimize innate character abilities, much like I2TO and Knave both do. Even if you don't want to run a game where equipment is all, it feels like maybe a good place to look for advice about making equipment matter more.

    2. So I read through the linked posts, and indeed they are fascinating, worth thinking about in the context of TTRPG design ... and I think wonderfully inapplicable.

      That is they highlight the fundamental differences between CRPGs and TTRPGs.

      TTRPGs don't seek replayability for example - which makes procedural generation usually an emergency stopgap or neat trick. Like if I'm running a world dungeon I may want procedural generation for when the players wander off the keyed map, but I don't want to run an entire improv campaign of procedurally generated spaces.

      I know it's hip for the current generation of ultra-light games, but to me it often feels like the disassociated monster zoo sort design/play that dungeon crawls always get accused of being. That is to me procedural content feels too video-gamey.

      Also hard agree on "would this work better is a computer did it?" I also like "Does this mechanic add to play enough to justify its cost in time and effort?" The answer to the second is usually a "no" for me.

    3. @Gus L -

      "TTRPGs don't seek replayability"

      I don't know about this. Isn't a megadungeon supposed to be replayable? Is that not part of the appeal? And they are often procedurally generated (in advance, not usually at the table).

      I agree with the "monster zoo" criticism, though it's never really bothered me. I love monster zoos!

    4. @Uncaring Cosmos

      While a megadungeon has aspects of replayability, and any published project can of course be replayed, I find emphasizing these similarities is less useful then interrogating the differences.

      With a CRPG replay is literal and likely frequent, and though increasingly (and maybe historically given Nethack's ghost levels)open to static/evolving worlds the concern for replayability is uniquely significant. I think it is the driver behind a lot of procedural generation even - making sure that the specifics of a level remain novel through multiple plays. The player of a CRPG or even more a roguelite (say Nuclear Throne) is expected to play through the same content many, many times and that play needs to vary to the point where it's interesting.

      Beyond the interesting question of when a TTRPG megadungeon becomes setting as opposed to a mere dungeon (and so transforms replayabilty into something entirely different), replayability in TTRPG adventures is always imperfect, as well as far less common. No one says "Well that's another TPK, let's run Tomb of Horrors again! 105th time's the charm!" Partially this is a design question, but I think more its the nature of play. Having a living GM means that even if you are running ToH a bunch of times description and even play elements won't be exactly replicated each time. I've run Anomalous Subsurface Environment several times, and each time as a GM things change - not because I'm not using the same content (and some of the players have been the same as well), but because as a GM I don't have the same fidelity to the product that a computer would - I can't.

      Obviously there's another discussion here about tournament modules, 5th edition adventure path tombs, boxed text, Adventurer's League and all the rest of the efforts to create a consistent play experience across tables ... but I think the (at least among those interested in older editions and playstyles)consensus is that these things detract from the play experience of TTRPGs. TTRPGS are also group/social entertainments which increases variability without much effort. I think this can be accomplished somewhat with multiplayer CRPGs (I know I ran the Wailing Caverns in WOW a lot of times with different groups in 2007) but it's inherent in TTRPGs.

      In a CRPG fidelity to a specific vision/scenario is a problem that has to be addressed either through changing content across plays or through a save system (you can have a linear story if there's no requirement to start at the beginning again each time something bad happens).

      There's a lot of differences here, and they highlight the strengths and weaknesses of each of the forms - TTRPG and CRPGs. I find examining them interesting, but I still think the lessons of one form don't apply very well to the other.

      Monster Zoos are a style of dungeon design. I think they work fine for tactical combat games, but badly for exploration games. Tactics games are a lot easier for CRPGs (crunch that a TTRPG table would find impossible!) while exploration plays to the strengths of the TTRPG (a living thinking and adapting GM/open palette for player solutions to obstacles).

    5. Oops! I posted my reply in the wrong place. It's further down the comment thread, sorry :-(

  5. Pretty good analysis. Tabletop games and video games have a lot to teach each other, even though they're often completely different things.

    I believe Jeremiah Reed has hit the nail on the head about identification in roguelikes. In any roguelike where there's a gazillon possible properties for an item, the identification minigame boils down to "cast Identify if you can, or use that thing and pray". You can't even know the odds of good vs bad stuff happening, unless you've read the game's wiki/source code.

    But if there's only ten, five, two possible options, suddenly the choice's a lot more manageable. If seven options are bad and three are good (or the other way around), suddenly you might have an idea of the stakes. If you've only handling a small number of possible outcomes, suddenly you can start strategizing for each.

    This is something I think small-scope roguelikes like Brogue and Infra Arcana get right and heavily benefit from, compared to the giants out there. In any case, I'm definitely stealing that "you know this item is either X or Y" idea!

    1. Yeah, normally I don't feel very interested in having unidentified items, but using the idea of listing 2-3 possibilities makes it seem more tempting.

  6. @Gus L

    Cheers for the reply - a really interesting discussion!

    I want to be very specific, though, I'm really only talking about roguelikes and early D&D. Your broader point about TTRPGs and CRPGs is sound, but I'm not talking about a comparison of all CRPGs with all TTRPGs. When I say "replayable" and "procedural generation", I also hope we both have the same thing in mind, because it's possible those terms mean something slightly different to both of us.

    I think of a dungeon being "replayable" if players keep coming back to that dungeon repeatedly and regularly throughout a campaign. It's basically a "campaign tent pole". And "procedural generation", for me, means that parts or all of the dungeon is created by rolling on random tables (whether it's wandering monster tables, or dungeon stocking, or even using Appendix A to design entire levels) rather than by hand.

    I suspect a lot of modern CRPGs emulate RPG modules, and so have a linear plot (even if it's hosted in a sandbox world) with a BBEG villain and an epic conclusion. However, I'm more interested in roguelikes because I think they emulate earlier (circa mid-to-late 1970s) D&D. That's sort of "pre-module" D&D (though, of course, there were definitely early modules around). Modules kind of codified how people were "supposed" to play D&D, and I don't think early CRPGs (the really early stuff coming out of university campuses) were emulating that kind of module-driven play. Early D&D was (not all of it, but certainly a fair amount of it) based around dungeons that were designed to be replayable and to have their content refreshed ("re-stocked") using procedural generation (i.e. random tables).

    "No one says 'Well that's another TPK, let's run Tomb of Horrors again! 105th time's the charm!'"

    That is absolutely true. But I can imagine people saying: "Let's start our 105th sortee into the dungeons beneath Castle Greyhawk, and let's see what's changed since we last visited".

    Early CRPGs began as an attempt to emulate the early D&D experience. The sort of "replayability" I'm talking about was baked into CRPGs from the start because it was also there in the early D&D experience.

    I really like the way Anne put it in a comment on my blog: "[CRPGs] are in a conversation with D&D, and alongside the Choose Your Own Adventure, Fighting Fantasy, and Dragon Warrior books, are an attempt to recreate the social experience of D&D for a lone player with a digital Dungeon Master."

    So, I do think interrogating early CRPGs is a fascinating way to shed light on early D&D, and people playing that style of D&D today can draw from it usefully.

    "Monster Zoos are a style of dungeon design. I think they work fine for tactical combat games, but badly for exploration games."

    I see what you're saying, but if you're playing in a monster zoo you do not actually have to fight the monsters. For me, monster zoos support exploration-focused games because they facilitate faction play. Having loads of different tribes of humanoids living together in the Caves of Chaos, for example, means you can Yojimbo the kobolds off against the orcs. This also applies with individual monsters: lure the ogre to the owlbear so they fight (or maybe they gang up on the party!), for example.

    I don't really care WHY the orcs, kobolds, owlbears, and ogres live within spitting distance of one another. I care that all these different monsters have different properties - wants, fears, strengths, weaknesses, strategies, etc. - and that emergent gameplay can come interacting with all that variety. I also think a lot of (very) early D&D took the monster zoo approach. For example, Castle Greyhawk, from what I've read, was a monster zoo.

    1. P.S. I guess my point could be summarised as:

      The OSR (or, at least, some of it) emulates early TTRPGs.

      The roguelike scene (or, at least, some of it) emulate early CRPGs.

      Early CRPGs emulated early TTRPGs.

      So, maybe the OSR can usefully pinch ideas from the roguelike scene (and vice-versa).

    2. @Uncaring Cosmos

      Indeed, great response.

      I tend not to talk about the "OSR" as I find the label both meaningless and placing too much meaning on certain bad elements of the hobby - I say that despite having been an active participant in its culture from 2011 - 2016, and I think a moderately well respected one. I do still think a little about "classic" TTRPG play and am well enough versed in its products and playstyle of the OD&D era.

      I think one important aspect of the early modules you cite, but I think Jaquay's early work, Tegal Manor, or Palace of the Vampire Queen might be a better illustration of the megadungeon or even exploration dungeon then the Gygax & Arneson early works.

      The point being that within even the earliest modules there's a lot of variation - which I tend to class into three sorts of dungeon adventures: the Seige (fortress with organized opposition - G1 is a great example, as is Temple of the Frog), the Dungeon (exploration of a location with many rooms of mixed puzzles, encounters and other obstacles - Thracia or PoVQ) and the Maze (a trap and puzzle focused adventure - like ToH). All of these sorts of adventures are found before 1979, and most are hybrids of the three.

      The core experience of a roguelike misses or has great difficulty with most of the gameplay styles in these adventures: puzzles, roleplaying/negotiation and faction play. Really the only elements a CRPG roguelike (and yes they've been getting better - ADOM isn't Rogue) manages well is combat and treasure discovery. The ur puzzle of dungeon adventures - spatial orienteering and mastery of the level map exists, but since few roguelikes expect that the PC will leave and return to the same level many times, it becomes less important. So yes, it's obvious that rogue and nethack and moria are trying to emulate OD&D, B/X, or AD&D more or less - but they largely fail. Where they succeed is in the surface setting and occasionally capturing the same mood or experience. Not to say they are bad, but where they succeed it seems to me to be when they make their own experience.

      Alternatively, if one looks at one of the best Roguelites of the past few years, Darkest Dungeon, it's obvious that the puzzle and roleplaying aspects (even spatial orienteering) have been largely discarded ... along with most of the other roguelike standards e.g. the loot/equipment cycle. Yet Darkest Dungeon, far more then many CRPGs manages to capture key elements of the classic play experience: supply, light, encumbrance and the entire risk v. reward, luck pushing experience. I think it does this precisely because it seeks a core classic TTRPG experience (pushing one's luck and risk management) and then figures how to manage it without emulation of TTRPG rules.

      These are just examples and I've gotten off track a bit.

      Of course there's been cross pollination between TTRPGS and CRPGS - first one way and then both ways. For me the question, and I only operate in the TTRPG space, is if it's often a fruitful pollination. More often then not I think it isn't. There's plenty to learn and concepts to share, but the problems of the mediums are very different, and the lessons I think are found more in the differences then the similarities.

      Also we are using fairly different definitions of replayability (I'm thinking 'multiple campaigns' or 'multiple parties' to your 'multiple expeditions')likewise I have a more specific definition of procedural generation - meaning the creation of whole locations or adventures, usually improvisationally as they are being run (which I'd think is how most rougelikes work) - and there are certainly interesting efforts in TTRPG land in this direction, though I find them unsatisfying to a degree as mentioned.

    3. @Gus L

      "I tend not to talk about the 'OSR'..."

      That's very fair. I often use it as shorthand in the absence of another satisfactory term (and because of familiarity). However, I agree it also has a lot of toxic baggage attached.

      "The point being that within even the earliest modules there's a lot of variation..."

      Again, good points - and I completely agree about there being lots of variety in how TTRPGs were played right from the very beginning. Everyone was making it up without much guidance or coordination, so it was a case of a hundred flowers blooming. I also think the Rythlondar Chronicles is a good illustration of the kind of game I'm talking about.

      "The core experience of a roguelike misses or has great difficulty with most of the gameplay styles in these adventures: puzzles, roleplaying/negotiation and faction play."

      I think that's only partly true. I was playing Caves of Qud just the other day and it has all of those things. What about factions in Dwarf Fortress Adventure Mode or Cataclysm: Dark Days Ahead? Or conversations in Ultima Ratio Regum? Or puzzles in Brogue or Nethack? True, they're not perfect. And, of course, we get into the question of what is a "roguelike" (but I'll put that aside for one moment).

      "Since few roguelikes expect that the PC will leave and return to the same level many times..."

      I'm not sure... I can definitely think of roguelikes where the player is expected to keep returning to the surface to heal and restock (definitely not in all roguelikes, and certainly some classic roguelikes have evolved their own tropes - like the player eating the bodies of the dead - to cope with an uninterupted descent into the underworld). Even the PLATO System game "DnD" from 1975 assumed players would be revisiting levels multiple times.

      "There's plenty to learn and concepts to share, but the problems of the mediums are very different, and the lessons I think are found more in the differences then the similarities."

      I agree they're very different mediums with different strengths and weaknesses. However, certainly CRPGs have learned (and continue to learn) a great deal from TTRPGs. Fundamentally, CRPGs were originally an attempt to capture the TTRPG experience (they've now become their own thing). That idea fascinates me, and I think looking at early CRPGs also helps shed a bit of light on how early TTRPGs might have been played (because the experience the designers were trying to transmit digitally was the D&D experience they themselves played). Early CRPGs are kind of like a digital time capsule in a sense (though it's possible to take the idea too far, given that technological constraints would have meant changing the experience to fit the medium).

      "Also we are using fairly different definitions of replayability (I'm thinking 'multiple campaigns' or 'multiple parties' to your 'multiple expeditions')"

      I can't say for certain (given I wasn't born until the 1980s) but I think Rythlondar Chronicles suggests the lines might have been a bit blurrier at the beginning (with multiple parties and multiple campaigns carrying out multiple expeditions in a static set of dungeons). I've also heard stories of tables of players each having their own dungeon and their own PC (or stable of PCs), and they would all take turns DMing in the same world. The table would decide, sometimes at the beginning of the session, whose dungeon the group would play that day. Everyone developed their own (replayable) dungeon in the same way players develop their PCs today. I have no idea how common that was (possibly not very common).

      "Likewise I have a more specific definition of procedural generation..."

      Yes, I would definitely say all random generation counts as procedural generation for me. Building something from tables instead of just deciding what goes where or running a module, for example.

    4. It seems like you two are having a productive conversation, and I don't want to derail it.

      I do want to chime in and say that procedural generation of the dungeon at the table, while the players are exploring it is something that definitely not everyone likes, but I rather enjoy it. I think "Gardens of Ynn" and "Stygian Library" have done more to popularize proc-gen during play than any previous versions.

      I also think that "faction play" and "talking to monsters" are two things you can do in a roleplaying game like D&D that are harder or much impossible in a CRPG, a gamebook, or a boardgame. Combat, obvs, is pretty easy for a computer to handle. The same with exploration and resource management. A CRPG or boardgame can probably handle resources just fine, and might even be able to spend more gametime on the experience.

      It's the social element that's hardest to reproduce. In a solo game, whatever kind, you're also most likely to tolerate being railroaded, in the same way you are while reading a book. It's only when you're playing with other people that you REALLY miss having meaningful choices.

      So there are things that I worry about adopting from solo games. I worry about adopting mechanics that are too hard to implement without a computer. And I worry about the loss of sociability at the table, and maybe even the loss of playing the role of a character rather than pushing a pawn around a game board.

      I feel like in any game, you have only so much time and so much attention to give. On some level, paying more attention to one thing means ignoring something else. One thing I like about procedural generation as a GM is that it puts me in the same head-space as my players, and all of us are focusing our attention on the sense of wonder and mystery of finding out what comes next together.

      I don't want to play a game where combat has taken over any other in-game activity. But tracking resources can demand quite a bit of attention too, especially if you WANT it to be a major focus of your game. I almost feel like a boardgame is a better venue for a game where characterization and conversation are intended to take a backseat to map-exploration and inventory-management. Those are going to be more meaningful if in-game time passes more quickly because you're not constantly stopping to have long conversations (or long combats). If too few "turns" pass during the session to allow for moments of scarcity or real RM dilemmas, then the attention that it soaks up is going to feel like wasting your already limited time.

    5. Brilliant comment, Anne!

      I couldn't fit my response in a comment, so I've written up a post as my reply:

    6. I'll go over to Uncaring Cosmos with a longer comment the gist is that Full Procedural Generation (that is a Gardens of Ynn style random generation of the entire location/adventure - and I should be making distinctions there)is almost always part of a Roguelike (maybe a necessity - certainly a core design principle), but for a TTRPG it's much more of a novelty.

      Resource tracking that is what I call "Supply" is one of the three pillars of exploration play, and yeah it's been tossed out by contemporary trad TTRPGs like 5E D&D precisely because it's perceived as finicky, anti-heroic and dull - but also because even early D&D never really had great rules for it. However - there's a lot of ways to fix that, largely with more gamified mechanics.

    7. Cheers, Gus! I look forward to your comment (and certainly agree that procedural generation at the table is a rarity in TTRPGs, though it is also fairly often the way I run my own games).

      "It's been tossed out by contemporary trad TTRPGs like 5E D&D precisely because it's perceived as finicky, anti-heroic and dull"

      I think the finickiness can managed (with more gamified mechanics, as you say)... The other two are actually reasons I really love resource management. I would say "mundane" instead of "dull", but "anti-heroic" is perfect. I love all things anti-heroic and mundane in TTRPGs, and try to bring that feeling out as much as possible in my campaigns (leavened with a healthy dose of black humour). A bunch of peasants starving to death in the dark is a great night (though it's never played straight. My aim is to never be adversarial, but to be on their side against the dice, and to have the players roaring with laughter at their losses, and really savouring their victories, not depressed and feeling powerless).

      "Even early D&D never really
      had great rules for it. However - there's a lot of ways to fix that, largely with more gamified mechanics."

      I've used the B/X rules fairly straight and they've worked pretty well, but I 100% agree that emphasising the abstracted, "game-y" side of resource management is a great approach.

    8. Uncaring Cosmos, thanks for moving the conversation over! Here's the link again for anyone following along: Resource Management, Roguelikes, and Role-Playing Games

      Gus, I know I push back on you a bit, but I do want to say that I think the Siege / Dungeon / Maze distinction is really interesting, and I really like your mission to create better RM mechanics.

      Uncaring Cosmos, your summary of the potential relationship between OSR and roguelikes was really great.

    9. Anne, never noticed anything but support for my quixotic classic games project from you - liking different play styles and theorizing on them as in opposition to classic dungeon crawl is totally cool - I enjoy reading it.

      For the Siege/Dungeon/Maze categorization it's of course squishy. Adventures bleed between categories. G1 - Steading of the Hill Giant Chief is an almost pure Siege scenario (though Temple of the Frog in Blackmoor is mostly as well and earlier). I reviewed it here:

      As for old Dungeons I stopped blogging at DoS before I got to Palace of the Vampire Queen or Caverns of Thracia which fit the bill and are early - though Vampire Queen is the earliest and a blank slate - much like the undertemple in Temple of the Frog. Minimally keyed and moving towards zoo territory. It's worth glancing (all one can do really) at if only for its use of player maps and the way its one line room descriptions try to create faction and complexity - even if they largely fail. B2-Keep on the Borderlands - which really gets into the faction conflict/RP game is a fine example, but it's a bit second generation though.

      Jaquay's is of course the best. Caverns is a gold standard and remains both playable and visionary.

      Tomb of Horrors is of course the proto, and still very playable Maze. My review is here:

      We might also add a 4th category - the adventure path (wish I had a term not freighted with disdain) pioneered by the Hickmans in Rhasia and the first Ravenloft?