Tuesday 16 July 2019

[BLOG] Noise or Signal? Further Thoughts on Creativity and Randomness

Art by Virginia Frances Sterrett

Creativity aid, not creativity replacement? We have been there before, and it is a slogan uniquely suited to describe a family of products designed to help GMs develop their own adventures. Random generators have been used to jog our imagination and come up with interesting new combinations ever since Ready Ref Sheets and the Dungeon Master’s Guide appendices; while random tables fell out of fashion between the mid-80s and the early 2000s, they have experienced a revival and they are going as strong as ever. Procedural generation is heavily featured outside tabletop RPGs, permeating the worlds of computer games from Elite to Minecraft (not to mention the deep, dark well of roguelikes).

Randomness, of course, produces no innate meaning, and leaves us to project our own onto it. A dungeon room with a “fire throne”, an “ogre taskmaster” and a “magic warhammer” is a hodgepodge of disparate elements; it takes human imagination to connect the dots and turn the random gibberish into something meaningful. (Perhaps the fire throne is a torture implement, the ogre is a jailer, and he has stolen an imprisoned dwarf’s weapon? Or we are in the hall of the fire giant king, the ogre is his underling, and he is guarding the king’s symbol of power?) Just like modules are a framework to run an actual adventure, random tables serve as the framework for the GM’s imagination. And just like modules, the eventual results should bear the personal mark of the GM, and, ultimately, the whole game group. This is how we co-create, and this is how the whole can be more than the sum of disparate parts.

If it is all so subjective and variable, can you actually review a collection of random tables? Can you actually tell a good table from a bad one? I have used a lot of random tables over the years, and have found that some have proven consistently useful, while others are barely ever touched. There are qualities which make certain tables more suitable to provoke the imagination. It has to do with the entries’ imaginative power – their capability to evoke images which can be spun into fantasy adventures.

To work their magic, we have to trust the tables enough to follow them somewhere. But they must take us someplace special – imaginary places of wonder and menace. A table that does not push us out of our current frame of mind is not a good creativity aid, because we are already there. D&D has a common language – of oak doors, dark corridors, pit traps, wizards and goblins and maybe beholders – which is intimately familiar even to people who do not play D&D. They are “tropes” (a horrid word embodied in that most horrid product of internerd autism, TVTropes). Good tables take us beyond the basics – it is still the same language, but a richer, deeper, more varied layer of it.

Art by Edward Coley Burne-Jones
Some of the imaginative power of random tables lies in the strength of individual idea kernels, but just as much hinges on the combination and juxtaposition of elements which fit together in ways which are not altogether comfortable. Creative tension – the shock of unexpected combinations and the images they create – is what takes the mind beyond the limits of routine thought patterns. Yet there is a limit to oddity, where it ceases to be meaningful. Square birds in purple sauce? These elements don’t fit into a coherent hole. There has to be a “bridging” moment where the pieces shift together, and create something new. “Serpents” and “gates” are both powerful images in their own right, laden with symbolic significance – but a serpent-gate? That is surely something more. A “serpent gate mirror”? Now we are getting there. However, we are also getting more specific, which may limit our options, and reduce us to obvious paths where potent images are diluted back to cliché.
Results open to interpretation are better than static and immutable ones. This lies at the heart of the “oracular” power of tables – they tell the truth, but the truth they tell is different from perspective to perspective. This is a tricky balance to achieve – specific enough to be powerful, general enough to fit many different situations – and just vague enough. Dreams are the classic go-to example (and indeed, the Surrealists had already discovered this, including the use of random generation to combine dream-images). The best tables can be reused again and again, because their results have a universal character. This does not mean generic. The “Ruins & Relics” table from Ready Ref Sheets, the random wilderness encounter charts in the Dungeon Masters Guide, or the very first “Locations (Overview)” table in the Tome of Adventure Design all have a strong personality them that influences their results. Indeed, “Ruins & Relics” is as core to the identity of the Wilderlands as the DMG charts to “the AD&D campaign”, and the ToAD table to Mythmere’s vision of “weird fantasy” as the key to the rediscovery of old-school gaming. These tables are foundational.

And finally, there is randomness. A totally random generator is just the noise of a thousand monkeys at a thousand typewriters. It may produce something good – but it won’t. A random generator whose results can be predicted, or which does not produce novelty, is superfluous: everyone possesses the ideas it produces by default. And there is a third, subtle distinction: while a kaleidoscope always produces something different, it always produces the same thing – a kaleidoscopic image. It is a powerful tool, but limited.


  1. I don't think it's very complicated - it comes down to the power of the images included. Red goblin / blue goblin / green goblin - that's bland as sliced bread. But a simple matrix with armor / dream / tentacle and half-eyed / daughter / orphan can get you somewhere.

    The Dungeon Dozen or the Black Hack have the right proportions, but there are plenty of others. Btw William Burroughs' cut-up technique (cut apart a page of text and rearrange it) aimed for something in this vein.

    Let's see where the above could go: square birds in purple sauce - you may find this on some monster's table, the birds from a bizarre world next door with square life forms or twisted dimensions. Serpent gate - perhaps just a portal carved in the image of a serpent. Or giant serpents turning into round interdimensional portals when biting their own tails. How can one coerce them to do this or decipher which kind of snake goes where? Or perhaps you just need to be swallowed?

  2. ==The “Ruins & Relics” table from Ready Ref Sheets, the random wilderness encounter charts in the Dungeon Masters Guide ... These tables are foundational.

    These tables are instructive and exemplary but not foundational. There is no imperative to build *on* or *from* them. IMO what most people don't understand about these fundamental random tables is that they are abstractions of and reductions down from a core of ideas that have already been established. Random tables should be a crystalization of immense knowledge. Typically since 1982, especially in the OSR random tables are lists mindlessly mashed together in the hope that sparks will fly. The original concept was 'how do we capture the immense knowledge we have on this subject in a way to create variety'. The OSR concept is 'here are diverse lists which when combined will create random, foolish or bizarre garbage.' This might be a reason why AD&D people form the 1980s don't recognize OSR writers.

    TLDR - good random tables condense or abstract knowledge; worthless random tables mush lists together in the hope of creating sparks.

    The first two JG Wilderlands supplements are jaw-dropping. The tables are unwieldy but magnificent. The reader always has the sense that the outcomes are contained.

    The AD&D wilderness encounter tables are exemplary in the sense that they condense information that Gygax had in his head, masses of data to tables of data. Poor random tables compound arbitrary lists, into a gibberish of hopeful explosions, that is not the way to create.

    1. If your idea of condensed wisdom is "200 orcs" or "3 pterodactyls", then sure. But there's a reason why people stopped using these tables: they are boring as hell.

      Other tables can produce gibberish, of course, but it can be avoided even with a tiny bit of common sense and intuition. If it sounds like artsy BS to you, it will probably sound like that for your players as well.

    2. 200 orcs AND 3 pterodactyls, however... Orcs vs. pterodactyls? Orcs riding pterodactyls? Intelligent pterodactyls ruling over savage orcs?

    3. It is not any individual entry that requires knowledge but the entire information content of the table. Idiots in the OSR collect things into tables which don't go together because they don't understand what they are trying to randomize.

      A good example is a random terrain generator. You need to understand geology to create a table which would produce sensible results. OSR idiots have never understood this principle.

    4. Maybe OSR idiots are not interested in geography, just in playing a game, and the tables favored by them reflect that, their very point being the juxtaposition of things that don't fit but may reveal ideas never seen before.

      Btw I doubt that even Gygax had the whole range of bona fide geographical, demographical, biological etc. credentials for such a serious undertaking. He just spew forth reams of tables that may _look_ "scientific" to eager little nerds, but in fact are just as arbitrary as the surreal output of OSR idiots, just way lengthier. I mean, we're discussing the distribution of sphynxes and manticores over imaginary mountains and wastes - how "scientific" that needs to be? And what use that "science" would have?

      Note that I'm not advocating total senseless randomization. Even imaginary worlds need cohesion for the players' actions to have real consequences and weight (or to put it another way, if anything is possible, then nothing is interesting). I just can't help but marvel at some people's dogged determination to drive a square peg in a round hole - i. e. enforcing utterly boring (and what's worse, almost always useless) "scientific" approach on something that serves as a vehicle to escape from it.

    5. You are missing the point. It's simple. Don't randomize something you don't understand. If you are knowledgable in a subject then your random tables will be interesting and intelligent and useful. Since Gygax created the milieu from scratch his random tables are definitive, instructive, interesting and useful.

      On the other hand, OSR people, all of them not most, don't appear to have any knowledge about any subject. It is as if they venerate the early-teen state of knowledge of the world, far into their forties, which is immature, illogical, ridiculous and pleasing only to jam and faeces smeared infants.

  3. Hey Gabor, I'd like to discuss publishing some of your posts in a new OS gaming magazine. Would you be able to send me an email (I trust my gmail address shows up in your moderation panel)?

    1. I can't find that info, but do mail me at beyond DOT fomalhaut AT gmail DOT com!