Saturday 4 November 2023

A Patterned Magic

Here's a mini-game for a wizard player, designed to encourage tinkering between sessions in a way that emulates downtime magical research.

Imagine a sudoku board with astrological symbols, which corresponds to the pattern of the wizard's preparations and meditations.

Certain tetris-like shapes of specific symbols are how wizards memorize spells. e.g. Fire over Earth with a Sun on both sides of the Earth is the Smiting Fire spell. If that's placed anywhere on the pattern between sessions, Smiting Fire is memorized.

Smiting Fire

Great Patterns

The GM prepares five or so Great Patterns, completed boards which each correspond to the ultimate expression of a particular divine imprint on the world, ancient unholy covenant, or the most secret teachings of a long-lost school of magic. In addition to the memorized spells, the more of the board that matches a specific Great Pattern adds different effects. (If you unwittingly include the shape for Drenzel's Eye, orcs will always find you, etc.)

There are also basic, sudoku-like rules. It's not good to have more than one of the same rune in any given 3x3 'house', for example. The the more rules you break, the higher the chance of spell failure or wacky outcome.

A small pattern containing 6 different spells

Newer wizards can always leave large parts of the pattern blank for safety's sake, or only bear smaller patterns less than full 9x9 size. Wizards bearing patterns with more than a dozen runes radiate faint magic to those who can see; wizards with completed, full-size patterns are psychically radiant, crackling with otherworldly sparks and eddies that tumble from their hair and fingers.


Wizards advance in a few ways: Fragments of these boards are scattered throughout the campaign world.. a strip of runes on an altar, a 3x3 embroidered on the lich's frayed shroud, a dusty tome of mad scrawlings.

Sometimes it's a literal spell, sometimes it's just a part of a Great Pattern, or perhaps just a clue. ("The great meditation of Deel wants not for fire in the upper houses, and thrice Water.")

A wizard's meditation pattern can be changed in play, but this is usually a solo activity done between sessions. (It takes a game hour of time to change one rune if it does happen in play.)

Another way to advance is magical research. If a wizard has a month of downtime, they can submit a pattern of whatever level of completion to the GM. They make a research roll, modified by the money they spend, the time and the quality of their sanctum.

Based on their research success, the GM will reveal one or more clues, complete spells, or matching parts of a Great Pattern in the wizard's submission. (This could be a useful discovery, or perhaps something to avoid at all costs.)

Another way to learn more magic is to watch another wizard perform it. When encountering either instructive, helpful, or antagonistic wizards using magic, there is a chance for the player wizard to discern a fragment of the meditation in use.

A Magical Minigame

The point of all this is to give wizard players a minigame that feels a little bit like actual magical research. Bring the arcane arrangement of runes and symbols into between session play.

The 'minigame' is the tinkering with the patterns, taking all the scraps and clues of what the party has learned and trying to make the best pattern they can. What if that altar was actually to Deel, and the strip of runes we saw there does belong in the void pattern? How do I sandwich all eleven spells I know without having all those pesky suns fry me alive every time I cast a spell? Do I fully commit to the Great Pattern of Deel, or do I try to bash in a few of those handy Sorgite curses in the bottom row?

Some Assembly Required

This of course only works if the GM is ready to prep a few things, giving runic patterns to all the spells, creating however many Great Patterns.

Most of all, it all relies on some tool to quickly enter and analyze a player-submitted pattern so the GM can immediately see everything of interest that will affect gameplay.


  1. Man, this is such a cool idea.

    How many effects do you think would be a good idea to include in a given pattern? Like 9?

    1. If you build the Great Patterns and your spell lists at the same time, you could 'cheat' by making the great pattern first, then just artificially circling parts of it to define as spells. For example, Deel the life-giver's 9x9 Great Pattern could be chopped into easily 20? spells if you let them overlap. (I'm imagining lesser magics as 3-4 symbols, and maybe greater magics are 5-6?)

      How many spells an adventuring wizard crams onto their board would depend on a) how much they know, and b) how clever they are at packing them in, and c) how well they understand which magics are related (from the same Great Pattern). Somebody trying to fit in a curse of Sorg from the Unholy Pattern might find it impossible to get more than 10 or 12 spells in (hypothetically speaking).

      I'm assuming here that players don't automatically get access to spells whose patterns they happen to have on their boards. If they were travelling around with Smiting Fire memorized, they get no benefit, but "hey wow, there's a spell here" would be a great result for a successful research downtime project.

  2. Man this sounds awesome.

    Have you had a chance to playtest this? I was chatting with my brother about it and some questions that popped into my head are:

    Is the full Pattern always 9x9? Or would it work to have larger or smaller patterns?

    For constructing the Pattern, should it follow the other rule for Sudoku? (ie: rows and columns should not contain more than 1 of the same symbol.) And should there be spells that fit into the Pattern, or should all spells be some type of rulebreakers? I could see an argument for either case.

    Also should spells 'consume' a given symbol during the day? Or should you be able to double-dip symbol usage between spells?

    What benefits should a completed pattern yield do you think? Would this be end-game type magic like *Wish*? Or just a really powerful passive effect, like certain types of magic just don't effect you?

    Thanks for making this post, it's really getting the creative juices flowing!

    1. Playtest? No, I only thought if it yesterday while making supper! I could imagine people only filling in a smaller pattern at first, perhaps 5x5? But I think if you're going to compare it to a Great Pattern it makes sense to assume you're seeing the middle 25 boxes of a sudoku board just so it's unambiguous

      I think the literal rules of sudoku might evoke the wrong thing (especially for Sudoku players), but perhaps any kind of similar-but-different constraints could be used. One danger with the specific constraint design is that a clever player might guess the entire board quite early? But perhaps that's it's own reward, and if spells have other limits (like the level-less spells in Wonder & Wickedness that grow in power with the player's level) that might just be part of the fun.

      I think the rewards for matching some or all of a Great Pattern should be pattern-specific. For example, Deel the Life-Giver's great pattern might give you healing, long life, immortality even? While Sorg the Devourer could be something else entirely.

  3. Ok you hooked me, I'm coding up a pattern-comparison tool. Are these sigils of your own making?

    1. Haha, awesome. Yeah, they're just hasty scribbles, not all the same size. I figured if there were perhaps 16 they could all go into a big paramedic string as 81 hex digits. How do I stay up to date on your comparator?

    2. Alright, I'm in. Here's the runes I would use:

      l - light
      s - shadow
      r - earth
      a - air
      w - water
      f - fire

      e - aether
      o - oblivion
      p - preservation
      d - displacement
      g - glamour
      t - destruction

      L - law
      B - balance
      C - chaos
      U - unity

      There are 100 individual spells in the Basic ruleset, if I'm not mistaken. That means only 10 9x9s with 10 spells each are needed. (Thankfully, that lines up with the number of deities in my setting lmfao.) Would you consider overlapping runes to count for each other? That is, if:

      . f .
      f d f
      . f .

      f . .
      . d .
      . . f

      If a player submitted
      f f .
      f d f
      . f f

      Would you consider them to have both Fireball and Firebolt prepared? Would using Firebolt expend the magic in those runes and make Fireball uncastable?

    3. I think overlapping is essential to make this work well, otherwise it's just a question of fitting in Tetris blocks.

      This system poses some challenges with B/X because of the Vancian 'fire then forget'. What if you want to memorize three magic missiles?

      B/X magic is powerful but limited by number of castings. I'm thinking of spells like Sleep that can wipe entire enemy groups at level one.

      For minimal modification to B/X, what I would say is that the your current pattern limits which spells you can memorize, but doesn't mean they are memorized. That makes it awkward, but someone used to B/X might find this layer acceptable in order to get the sudoku fun and the chance to learn new spells through in-play clues.

      But my sense is that this would work best with a magic system that's akin to equipment trade-offs. Plate armor makes you hard to hurt, but you're loud and swimming is totally out, and on its own doesn't break the game. Similarly, you can carry Evernard's Flamesnake in your mind and use it however often you like, but it's an extremely specific effect with significant drawbacks.

      I think this means that effects like Fireball would be /much/ higher level, something you'd only acquire after matching a huge portion of the Great Pattern of Infernal Unleashings or whatever.

    4. I said elsewhere that I think this system may work well with Wonder & Wickedness, which is a set of level-independent spells. Spells have no inherent level, but the level of the magic-user increases their effects.

    5. you Said you were coding up a tool, are you perhaps sharing the tool

    6. For the making and comparison of Patterns:

      Pure Python, I had plans for a friendlier GUI but got distracted

  4. This is a stellar idea. I'd pin it down to 9 elements/schools/alignments/etc. and keep all sudoku rules. MUs start with a 3x3 grid and 1d6 cells to fill. They can research for a certain spell(s). As they level up, they gain columns/rows and additional d6 cells to fill. The aim is to 'pack' the symbols to make as many *overlapping* patterns as one can...

    Spell A

    Spell B

    Spell C



    Note the need to separate the ☠️s and 💧s by 3x3 box. Could make for interesting combinations and incompatibilities! Every MU would come out differently.

    1. Something to bear in mind is that sudoku puzzles are fully resolvable between 17 and I think 25 clues. By level 5 or so, you might find yourself straddling an awkward line between giving the player clues they've already worked out, or giving a clue that unlocks the whole puzzle!

    2. This could gel well with your two other constraints: you can violate the sudoku rules, but every violation causes a "penalty" of some deleterious effect, AND matches to some Great Patterns are not entirely positive either

    3. Totally. I guess in my head it's more about spells known than prepared/memorized. (I recognize that's not what you were originally thinking. Haha.) I can imagine the whole 9x9 not actually solved, or even solvable, maybe even have a 9x9 but with maybe 12 symbols just keeping the no repeats in boxes/rows/columns restriction? Rather the aim is just providing a set of restrictions representing compatibility (or not) between different spells or types. I used to use alignment as an abstraction for this, in a sense, until it got gorked somehow.

      Now I'm also thinking different types of spells can be different layouts. Like combat or defensive spells are mostly columns, more conjuration and divination in rows, more complex effects are more complex shapes... My gears, they are a turning.