#dev pipeline

When I ran the Red Dragon's Tale, I took advantage of the dollhouse play features to present encounters in isolation, and used tissue paper to conceal chambers that would not be within their character's diegetic line of sight. In doing so, I was able to tell my players, "what you see is what you get". I invited them to use flashlights, magnifying lenses, and offered to disassemble portions of the model to let them have a closer look – in all cases, their real-world senses dictated their awareness of the fictional environment.

This approach eliminates the need for perception checks, which are a perpetual flaw with most mainstream RPGs. Perception checks lack obvious consequences for failure (meaning that they are unable to alter the game state unless they succeed) – except for outright denial of information, which all-too-easily stalls the game when the players lose access to details that they need in order to progress.

But prior to the invention of the perception check, the standard approach was for the GM to simply decide what the players can and cannot perceive – a sort of Captain May I? gameplay. It's so difficult for most folks to thread the needle between things being blatantly obvious and things being impossible to find that some OSR designers advocate for avoiding hidden information altogether. Hell, Robin Laws' entire career from about 2007 to 2012 consisted of just publishing this single idea over and over again.

The majesty of the Lego model is that it lets you hide things the way you might actually hide them: camouflaging them, putting a conspicuous box in the way, concealing them in plain sight – without any need for mediation by either dice or the GM during play. Once constructed, the model simply is what it is.

And that same simplicity extends to tons of other forms of interactions. A player asked if his character could keep his footing on a sloped roof – so I simply offered that if he could find a stud or bar to anchor his minifig, then he wouldn't fall. When another player triggered a pit trap, we got to physically see the minifig fall through the hatch. When a player spotted a potion through a crack in the wall, we tested whether his minifig's arm could fit through the gap to grasp the bottle. Whereas chance or GM fiat might have ruled in another game, here the physical realities of the model felt more fair, more "true".

It was an easy choice from here to make the game use strictly player-facing dice rolls. Each player would customize their own die, but the GM would never roll any dice at all – and so I'd never have to answer concerns about whether the GM or NPCs also get custom dice. This would also mean that I only need to actually procure a handful of the things, plus adequate tiles to cover the faces.

I hemmed and hawed at length about how best to exploit the customization angle, but the common thread between them all was quite simple: each of the game's character attributes would have a corresponding colour of Lego tile. Say, for the sake of example that the "strength" attribute is represented by the colour orange. So when the GM calls for a strength check, then the player rolls their die, and if they have an orange tile showing on the rolled face, they succeed – and otherwise they fail.

In this way, the die would functionally replace the character sheet: rather than knowing that your character has high strength because of a high-value number recorded on their sheet, you'd recognize a high-strength character by the fact that most of the die faces have orange tiles.

I toyed with tons of variations on this idea that I don't really have space to describe here, but I ultimately landed on just two major alterations from that basic premise:

First, one face on each die would be completely taken up with a single 2-by-2 white tile. This would serve a dual purpose: providing a space to write initials (or another distinguishing mark) that could be used to identify the owner of the die – and would also count as an "automatic success" result, thus ensuring a minimum 1/6 probability of success even if the player has no other matching tiles.

From a purely consumerist standpoint, this was an extremely good idea. Two of the 3700 bricks were misshapen, and several unique specialty parts (magic swords, armour, hairstyles, etc) were substituted with default alternatives – but at about 1/5th the price point (before shipping), it's still an overwhelmingly better purchase than the official Lego kit. I can't speak to the labour practices or other ethical concerns, but the knowledge that I might be able to cheaply obtain large quantities of nearly-perfect bricks will be important later in this story.

And yes, this post is something of a product review, but I promise all the details will also bear some relevance to the development process.

If you're into Lego-as-display-piece, it must be remarked that the kit looks amazing – but I'm not, and as a toy, it's quite underwhelming. Except for the largely empty meadow and the totally empty bridge, all of the playable spaces are extremely crowded, cramped, and awkward to manipulate. Only a few of them disassemble in ways that offer better access, and they are almost all constructed to be accessed from the side (like a dollhouse), which means that foreground objects obstruct background ones, and it's hard to apply the downward pressure needed to anchor a figure to the floor studs.

It's not an accident, apparently, that traditional dungeon crawl games used a top-down perspective for mapping. Thanks to gravity, most objects rest against the floor of a dungeon, and relatively few things are suspended in the air – so foreground occlusion is less frequent if you view the dungeon from above. Regardless, I also made note of the fact that spaces in a physical model need to be large enough to accommodate fingers and hands in between the walls, furniture, and figurines.

Finally, the kit also includes a disastrous, plot-hole-ridden, and debatably-broken D&D module authored by Chris Perkins that reads like a rushed first draft. In the process of completely rewriting it for my usual play group, I also converted it for use with Into the Odd, since I can't stand to play D&D anymore. If anyone is interested in that, send me an ask. The results of this little playtest were interesting: while the group agreed with my various criticisms above, they all felt that the physical interactivity of the dungeon module held incredible potential – and that a creating a purpose-built dungeon and system would be a very worthwhile endeavour.

I have to give all credit to Vincent and Meg Baker for providing this incredible design chestnut (emphasis mine):

Over the years, Meg and I came to pay close attention to one particular way to assign the say. It was Meg’s default for running lightweight games with kids, for instance. All things equal, it reliably promoted enthusiastic, collaborative play, and kept people engaged through the highs and lows of play. If you’d asked us, we would have recommended it as the simple best choice, for any game that didn’t demand differently: The GM describes your successes, you describe your failures.

Laid out like that, it seems obvious: giving a kid the opportunity to describe their own failure states means that they retain a sense of agency even when things aren't going their way – it's not an abject loss of control. Instead, the player can take ownership of the failure state, which provides both consolation and optimism as they reframe the failure in terms of the opportunities that yet remain open to them. I lifted the idea wholesale.

I was also anticipating that major commitment to the game would be in short supply, so I decided that the game should skew towards a "one-shot" format rather than long term campaigns. To that end, I created short-term advancement mechanics: during character creation, a player would need to describe a key personality flaw and a personal goal of their own character. These elements would be represented by 1-by-2 tiles blocking some of the customizable studs of their die.

Rolling the "goal" tile counts as an automatic success on any check if the player declares that the action is in service of their goal – but when so used, it then has to be replaced with normal attribute tiles. (It can also be replaced with attribute tiles whenever the player decides that their character's goal has been achieved.) The primary function of the goal die is actually just to provide the GM with some fodder for story ideas – giving a sense of what the player wants to see their character do.

At the time, I was running our school's D&D club with essentially no budget – a dozen members sharing a single copy of the rulebooks and using spare school supplies for character sheets and maps. We tracked character positions by writing their initials into grid paper cells; erasing and rewriting them when they moved. My point being, nobody in that club was about to invest in a Lego collection big enough to facilitate our adventures, so although it was a cool idea, it was out of the question.

Fast forward to 2024, and Lego and Hasbro release the cross-promotional "Red Dragon's Tale", a 3700-piece kit with a price tag in the hundreds of dollars. My RPG-disdaining-but-Lego-loving brother was quick to forward the promotional materials to me – but I staunchly insisted that I was not about to reward the Lego corporation with such an obscene amount of money in exchange for this obvious nostalgia bait.

Nostalgia is a potent thing. I dismissed the absurd Lego kit itself, but all the same I started to think back to that Lego RPG with the "backpack" box. I still didn't want to invest in a Lego collection – and my brother, despite already possessing a vast collection, would not play an RPG... but it dawned on me that my 6-year-old niece has both a passion for games of pretend, and a growing brick collection of her own (funded, of course, by my brother).

Now, I would never dream of inflicting such an awful RPG system as D&D upon my own niece. But I thought that perhaps I could design a new system, one that she would be able to grasp and that would integrate the Lego toys directly.

I filled the Blackstone forest with as much optional material as I thought could fit while still keeping the game relatively “lite” in terms of the logistical demands on the DM and the overall scope and duration of the adventure. Everything that exists in the forest somehow derives from, or feeds back into, the central quest, but there are a lot of incidental stories taking place in the forest that don’t involve the player characters (PCs) – until they stumble into them. The vibe I wanted to achieve was to hand the players an initially vague objective, a time limit, 5600 square miles of exploitable resources and let them sort the rest out on their own. 

Accomplishing that goal meant, in part, shifting the incentives around a whole bunch. One of the earliest decisions I made was to set and lock the PC level at 6. This would provide the players with some options for their builds (expanded by the inclusion of background utility powers and a feat tax subsidy) while keeping them far away from the power creep inflation, removing XP as the default overriding incentive, obviating the need to calibrate encounters for a range of PC levels, and increasing the incentive focus on magic items and consumables. 

Magic items that, since I banned the enchant magic item ritual, can only be obtained by hunting down monsters and harvesting their bits. I kept this system really rudimentary, but if my theory is correct, it should result in some entirely player-driven side quests to outfit themselves by tracking dangerous prey. And, even if it fails utterly, there’s enough failsafe built in that the PCs will still have adequate gear to see the main quest through (although if it totally bombs I’ll likely nuke the entire crafting system before publishing). 

The design is that no matter what the players do, they’ll eventually get some gear, learn the clues, solve the mystery, and show up like Big Damn Heroes to save the day before time runs out. The gamble is to see if the forest can be made into an interesting enough place that the players will take it upon themselves to go exploring in it for its own sake: they don’t technically have to do it to win, but it costs them nothing and yields high adventure. 

One of the original design goals of Blackstone was that it should contain some kind of “mystery”. That stated goal was fairly vague, but I take it to mean that it should be possible to rapidly or easily resolve the major conflict of the story – if (and only if) the players are privy to some key piece of secret information. The act of winning the scenario is almost trivial, but figuring out how to win is hard. Once the mystery is solved, victory is a forgone conclusion, but solving the mystery requires the players to obtain clues from which to deduce the solution. 

Now, the best way to play out a mystery “realistically” is to be thorough and detailed in your conception of the facts, and then try to objectively present the logically available evidence as the players ask for it. This is a bit of a nightmare though, since you and your players can very easily be on totally different wavelengths, and it’s also very difficult to walk the tightrope between simply giving the solution away and utterly burying it under obfuscation. 

In exchange for a significant lapse in plausibility, however, you can instead construct the mystery as a puzzle of deductive logic delivered as a set of self-contained clues, so that the players can deduce an unambiguously correct solution (a notion that occurred to me almost exclusively because of how The Consuming Shadow does this exact thing). This allows you to arbitrarily define the quantity of clues required to solve the mystery (by adjusting the quantity of variables included in the puzzle) and therefore determine the effective duration of the investigation. It also means that you don’t have to worry about the order in which the players receive the clues, or about any one clue being more or less vital than another, so you can basically treat clues like pieces of loot to be handed out here and there as part of the normal process of the adventure. 

This is mostly because 4th Edition is very deliberately designed around the concept of an “adventuring day”, where the player characters (PCs) completely regenerate all of their resources any time that they go to sleep. They heal all their injuries, recharge all of their powers, and (thanks to starting-level, commonplace magic items like the “everlasting provisions”) even renew their supplies of food and water. This pacing works pretty well for the typical dungeon-plumbing adventure for which the rules were designed, but it’s useless for an adventure that hinges on long-term attrition. 

To introduce scarcity, I first dropped a general-purpose ban on sources of unlimited food, water, or shelter, so that the PCs would be forced to carry stocks of each (which would, in turn, dwindle and periodically force them to obtain more). Then I added a special restriction to the recovery rules: resting doesn’t work unless the PCs do it at a designated safe haven – this way, the PCs can spend days on end out in the forest, and in mechanical terms it’ll be no different than if they spent a few hours in a dungeon (except, of course, for the rates of food and water consumption, which are now more comparable to the rates of health and power consumption). 

From there, I added in a few extra survival systems to make the non-combat elements of the adventure more robust: 

deployed 4E’s general-purpose skill challenge mechanics to create rules for foraging for food and supplies. 

brought the rules for hypothermia, starvation, thirst, and fatigue together under a single “exhaustion” mechanic so that they'd all be consistent and would be balanced for Blackstone’s pacing. 

reworked overland travel so it functions like a scaled-up version of the game’s combat movement.

invented hotter/colder-style mechanic for finding points of interest in the forest and tracking down suspects