A Rasp of Sand (“AROS”) is an adventure written, laid out and illustrated by Dave Cox and was funded as part of the original “ZineQuest” for Kickstarter. At 75 pages, AROS pushes against the definition of “zine” and may be more appropriately dubbed a “book.” AROS is written at times with all OSR systems in mind, but it is so thoroughly integrated into Knave by Ben Milton (of Questing Beast fame) that it is more accurate to label AROS as an adventure for Knave, that could be run for non-Knave systems with some referee-applied elbow grease. AROS is a procedural dungeon crawl that is unique in its oceanic theming. But what really sets AROS apart is its ambitious goal: AROS explicitly draws from the rogue-lite subgenre of video games. AROS succeeds with flying colors in this endeavor, producing a cohesive adventure in the rogue-lite vein that encourages player mastery in a way that is unique from a typical dungeon crawl.
Disclosure: This review is not based on my experience playing in or running AROS. Instead, my conclusions are reached based on a thorough reading of AROS. However, I did also consult with @qpop in writing this review, who has run the adventure to completion twice for a group of 4-6 players over four generations (of player-characters, not of players—more on that later) and just ran a speedrun of the adventure with some of the same players! You can see a rundown of the speedrun in this tweet thread.
What is a Rogue-lite?
To understand why AROS makes the decisions it does, it is important to understand the rogue-lite video game subgenre. To answer what a rogue-lite video game is, you must first unravel what a roguelike video game is. The simplest definition is that it is any video game that is like Rogue, a 1980 ASCII-based video game inspired by Dungeons & Dragons. But truly unraveling the meaning of roguelike or rogue-lite would be a poor use of your time on a TTRPG review site. (Although, Anne of DIY & Dragons [and of here!] put together an interesting list of roguelike advice as it relates to and is helpful for tabletop gaming. It is worth checking out if you are interested in exploring this intersection further.) “Roguelike” is a hotly contested term. There is a “Berlin Interpretation” that sets out a number of factors, but such interpretation is far from being widely accepted. To skip past the quasi-medieval scholasticism surrounding these key terms, the two most important aspects of a roguelike are permadeath—when the character dies, they start the game over in a new “run”—and procedural dungeon generation. Rogue-lites shed the strictures of the Berlin Interpretation and embrace metaplay, allowing players to unlock persistent features that carry over to future runs. The rogue-lite to which AROS shares the most DNA is Rogue Legacy, a 2013 rogue-lite. The similarities between AROS and Rogue Legacy are mechanical and fictional and hinge on the idea of “heirs.” When you die in Rogue Legacy, you pick between three randomly generated heirs, which have chances to inherit characteristics from your now-deceased character. AROS handles it somewhat differently, but the idea is the same.
Heirs & Heritage
In AROS, you do not play a character; you play a family. Each run through the dungeon is performed by a generation of characters, called Heirs. This is the fictional justification for the progression between runs. Because characters are actually expected to die and allow the next crop of Heirs to take a run at the dungeon, AROS produces a truly deadly experience, unlike other games where the characters are simply too precious to die. The most important thing each Heir passes down to their successor is their memories of the dungeon, but this aspect merits its own section (see Memory, Metagaming & Mastery below). Heirs also inherit ⅙ of their predecessor’s XP, one heirloom and any mutations from their predecessors. [Practical note: according to qpop, the XP passed down was fairly insignificant, but the passing down of heirlooms was a driver of strategy, causing players to debate over who should carry what items, which ideally should be spread out across the party in case of wipeouts.] Mutations are special abilities that Heirs may gain when they consume the vital essence (or “Sand”) left behind by different monsters. AROS also uses a process tailormade to Knave for determining a new Heir’s ability defense and ability bonus based on the previous Heir’s ability defenses. Each family has a family trade, and Heirs gain special abilities related to these trades. It functions like a traditional class system (which Knave does not have), but new abilities are gained by new generations instead of by leveling up.
The family trade is interesting in that it provides a link between the families and their island community, but it is less effective at passing forward progress from one run to the next. For each of the other aspects passed down, the benefit varies based on what the predecessor did—how much experience they earned, what items they gained, which abilities they increased and which mutations they gained. By contrast, each new generation in the same trade gains the next ability on the list for the trade. It does not matter how successful the previous Heir was or how they used the abilities from their trade, a new Heir advances just by virtue of being the next generation. One of Knave’s strengths is its simplicity, and this reduces the simplicity of Knave without adding much to the feeling of progression. If I were to run this adventure, I would probably ignore the trade abilities, although I would still have players roll for the trade that their family performs and start with the equipment listed for such trade. However, qpop disagrees and said that, in his experience with the adventure, the trade abilities provided a nice sense of progression and a through-line for the Heirs from generation to generation. Some of the abilities, like the Academic’s fourth-generation ability to siphon XP from treasure were game-changing (it seems that the fourth generation ability for each trade is quite an increase in power for the Heirs, in general). My main quibble is perhaps that acquiring the next generation’s ability is too automatic. Imagine a generation of Heirs that dies in the first room (which apparently did happen in one instance for qpop and his players—a reminder that this is a deadly adventure). In such an instance, the next set of Heirs would get a new ability despite the unequivocal failure of their forebears. I prefer progression to reward some type of success; if it is automatic, it feels unearned, which I don’t love when I am player-side.
Permadeath, Procedural Generation & Repetition
For AROS to include permadeath, that core aspect of a roguelike (or rogue-lite), it has to function differently than it would in a single-player video game experience. Dungeons & Dragons and its descendants all feature permadeath in a certain sense—when a character dies, the player must make a new character (absent a method of revival, such as magic). However, this new character typically joins the existing party, slotting into the campaign where the dearly departed character left off. Even if the players experience a total party kill (“TPK”), there is typically no rewinding of the clock, starting the adventure over again from the beginning. Just as strict time records are kept, time inevitably marches forward.
When a Heir dies in AROS’ dungeon, the Deep Queen’s Temple, the dungeon will flood, forcing all surviving characters to evacuate or die. Upon returning to their island home, a generation of time passes before the next set of Heirs will enter the Deep Queen’s Temple. This means that any character death acts as a TPK. But it is the fiction of AROS that causes this to resemble permadeath from a roguelike. Namely, the adventure hook. Essentially, the Heirs are the descendants of people that stole the crown of the mightiest of all spirits, the Deep Queen, and created a lake, which drowned the Deep Queen’s son, the Green Prince. As punishment, the Deep Queen flooded the world. The people now live on the few islands that were once mountaintops. Once in a generation, you have a chance to enter the Deep Queen’s Temple, return the crown and set the world right again. This is not just lore; it’s lore with a purpose. This adventure is not designed for sandbox or self-directed play. Players will be myopically focused on getting to the end of the Deep Queen’s Temple, which will necessitate multiple delves from multiple generations of heroes.
Most adventures are single-use. Or to the extent they aren’t, they still aren’t intended to run for the same players over and over. White Plume Mountain is still fun to play through a second time, but much is lost when you know what is coming next. The Deep Queen’s Temple, on the other hand, is built for replayability (as is AROS itself). It accomplishes this the same way its roguelike predecessors do: procedurally generated dungeons.
A procedurally generated dungeon is not the same as a randomly generated dungeon. A randomly generated dungeon would be stitching together random parts, without rhyme or reason beyond probabilities. If you use the generators in the back of the AD&D Dungeon Master’s Guide, that is close to random generation. If a dungeon level were generated entirely randomly during each run, what the players learn would have no bearing on future runs. Procedural generation acts as a set of instructions to guide the randomness toward certain types of results. It is randomness with its own set of logic. Going through a procedurally generated dungeon allows players to learn about how the dungeon works, which may be useful in future runs. This stands in contrast to a classic dungeon crawl where the layout and rooms do not change. There, players learn much more about the dungeon. But each repeated delve is a less fresh experience. Procedurally generated dungeons are designed to encourage repeated delves by rewarding dungeon mastery (by the players, not the proverbial dungeon-master) while offering a new experience each time.
To produce a dungeon for players to navigate, the referee must navigate the dungeon generation procedures in AROS. The Deep Queen’s Temple has five levels, each reflecting a deeper level of the ocean: Pelagic, Reef, Kelp Forest, Twilight and the Trench. For each run, each level requires that the players must explore [3d10, keep the highest] rooms before the exit appears to them. For each room, the referee rolls for a location and a situation. There is a d20 table for situations (with 14 results), and each level of the dungeon has a d12 table for locations. That’s 168 possible combinations for each level. While players are likely to encounter the same rooms across different runs, the situation is likely to differ. To pick a slight nit, I would have preferred for level-specific situation tables, both to add variety between the levels and also to minimize flipping pages when rolling up rooms on the fly. At first glance, I greatly disliked the rule about players needing to explore a minimum number of rooms and then the exit will “show up.” But at the back of AROS, there is an example of mapping a dungeon level that suggests the level should be mapped out ahead of time, which mitigated my concerns about it. However, some trade abilities, like the Sailor’s fourth generation ability (it’s always the fourth) allow the players to influence the dungeon generation, suggesting that the dungeon is generated on the fly. Per the example in the back, AROS doesn’t require a grind through a certain amount of rooms. Instead, there is some looping, and the exit can be reached without going through every room. My advice to anyone running this is simply to use Shifting Sands, a fan-made tool that is a very impressive distillation of AROS, including Heir generation and dungeon generation. Shifting Sands is a truly excellent resource. While AROS is completely functional without it, this is a tool that would truly make it a breeze for the referee.
AROS eschews goblins, dragons and all the familiar beasties of bog-standard fantasy RPGs. In their place, AROS offers 42 new monsters to find in the Deep Queen’s Temple (the section, entitled “Creatures of the Deep” is curiously omitted from the table of contents, but this is just another nit). The monsters are statted for Knave but could be converted with a modicum of effort. Almost every monster also has one of the aforementioned Mutations. So if a character eats the Sand of a Dolpod (a dolphinesque humanoid), you have a 25% chance that your forehead expands to make room for resonating space you need in your sinuses and you gain the power of echolocation when you make loud clicking noises. My favorite mutations are of this variety, where the power is telegraphed by physical changes in the Heir. Gaining enough mutations may also impact what happens the Deep Queen is encountered (though for exactly how, you will need to play in or read the adventure for yourself!). There are plenty of excellent monsters listed in AROS to populate the repeated dungeon runs. One of my favorites is the Marionetta Squid, which has a gimmick that the picture alone easily communicates.
Using all new monsters is purposeful for the design of AROS. The monsters not only mesh with the adventure aesthetically—the monsters share an aquatic theme, and not in the lazy way of many D&D aquatic monsters (I have in mind Lacedons and Koalinths)—but also are integral to make repeated excursions into the dungeon a meaningful and engaging experience. When you fight a goblin in D&D, you probably already know what you are getting into. Even if goblins are new to your character, you as a player know their essential characteristics. But how much do you know about a chum? Could you, for instance, distinguish a balloon chum from a tropical chum? Unless you are a veteran of AROS, I suspect the answer is a definitive “no.” In AROS, the Heirs and the players themselves each learn more and more about the ecology of the Deep Queen’s Temple with each run. Because of the “Sand” in AROS, players are encouraged to metagame with this knowledge, as it represents something subsequent Heirs learn from their forebears.
Memory, Metagaming & Mastery
In AROS, when any creature, including a Heir, dies, they leave behind grains of sand, appropriately called “Sand.” When a Heir ingests the Sand, they gain some XP and the memories of that creature. This is how Heirs pass down memories from generation to generation. As AROS explains, “There aren’t any actual rules for this, at least not in the traditional sense. Essentially you can metagame with knowledge from previous runs.” I’m glad that AROS uses the term metagaming if only to allow me to talk about it briefly.
Metagaming is bandied about as a pejorative term and it is unduly maligned in roleplaying games. Metagaming is an acknowledgement that there is a disconnect between the knowledge of the player and the knowledge of their character as well as a taboo against using exclusively player-knowledge in games. An example in the above-linked Papers & Pencils blog post is the player knowing the effect of fires on trolls while the character has not experienced a troll. The taboo arises in more thespian-oriented games and is verboten because it is an example of breaking character. In the Wikipedia article on the subject, it says that metagaming is “considered unsporting or cheating”, although this claim is both in passive voice (considered by whom?) and followed by a .
While I agree with my colleague, Nick LS Whelan, on having no qualms with metagaming, AROS is intentionally designed to (1) encourage metagaming-like behavior and (2) justify metagaming such that player knowledge IS character knowledge, for the most part. Because there are no goblins or other recognizable trappings of fantasy RPGs, players and their Heirs start their first run through the dungeon with the same minimal set of knowledge. What is learned about the dungeon, from its contents to its denizens, is passed down from Heir to Heir. There is little that the player knows that their current Heir shouldn’t know equally well. However, while this is not metagaming from a definitional standpoint, it facilitates player-challenge-based play, which is a strength of metagaming. In TTRPGs, typically both the character and the player are being challenged in different ways. While the character may become better equipped to face their challenges through level-ups, finding items or other in-game processes of advancement, the player only has one avenue for advancement: becoming a better player. Mastery of the system, mastery of their character, mastery of the game, all of these improve over time. With repeated runs of the same (or similar, being generated by the same procedure) dungeon, the player gains knowledge that helps them improve at facing the challenges therein. Even outside of AROS, this is fun for some players and groups. What AROS does, and it is a great strength of the adventure, is that it bakes this type of player-challenge-based play into the adventure fictionally. It truly is not metagaming to use the player-knowledge from previous runs in a current or future run through the Deep Queen’s Temple because your Heir has the same knowledge, passed down from the Sand. This fictional conceit is a permission structure for players to embrace player-challenge and dungeon-mastery (of a different sort than is typically meant by “dungeon master”), free from any character-breaking that is considered (presumably, by someone) cheating. In AROS, it would be unsporting to not metagame (and, given AROS’ deadliness, self-sabotaging).
Press F for Action Prompts
While AROS ingeniously integrates its video game inspirations into TTRPG form when it comes to roguelites, there are small portions that feel video-gamey in a less-than-stellar way. This is not a problem unique to AROS; I’ve come across it so often that I have given it a name: the Action Prompt Problem. Action prompts are moments in those quick time events commonly used in video games. They allow a scintilla of player input during largely cinematic sequences, for instance, allowing the player to pay respects during a cutscene of a funeral by pressing a button. In TTRPGs, this impulse manifests itself in referee-facing cues that trigger based on player action that the player is unlikely to actually perform unprompted. Secret information is good, but players should have some reasonable inkling that the secret information is there or a reason to interact with the world in a way that might reveal it. I will go over one example of the Action Prompt Problem in action in AROS, along with a simple “fix” for it. Hopefully this example better illustrates what this problem is, why it is a problem and how to avoid it in writing site-based adventures.
The weeping statue. In one of the rooms of the Kelp Forest, there is an alcove with a statue of a weeping man, among other things. The referee-facing room description says “If the Heirs wash their face in the tears of the male statue, they can see a path across the abyss [the major feature of the room]. It twists and undulates but provides solid footing.” My issue is simply why would the players wash their face? Investigate the crying statue, sure, but nothing about the statue or the room indicates that this is the time for a skincare routine. Contrast this with the other statue in the room, which the read-aloud text describes as “a statue of a woman holding a small bundle wrapped in seaweed up to her chest. She similarly has water running down her face, but it pools onto the floor and runs into the abyss.” This statue has hidden information (as opposed to secret information; for a taxonomy, check out this excellent post from DIY & Dragons), but it is gleaned simply by “investigating the bundle in the arms of the female statue,” an action thoughtful players are already likely to take. My fix to the face washing prompt is simple: get the players dirty. Either through an ink cloud of one of the creatures that emits it, or have another enemy fling mud at them. When the dirty PCs come to the room with the statues, they do not have a reason to know that either statue has magic face-washing properties, but they at least have a potential reason to want to wash their face. Another solution is just to tell the players, if they so inquire, that the water coming from the male statue looks different than the water from the female statue. But I like the dirty face idea more. Sometimes it is fun when players push a big red button by dumb luck.
Despite a few pedantic nit picks (solely in an attempt to live up to my series’ name, I assure you), AROS is an unusually coherent adventure. That is not to say that other adventures are incoherent in the sense that they are incomprehensible; I mean to say that AROS coheres together to form a unified whole, while I approach other adventurers like a car mechanic in a junkyard, looking for parts to rip out and insert into whatever I’m currently building. I am not sure that there are any specific, discrete parts of AROS I would steal, because each part is so codependent on the rest. AROS is the rare sort of adventure that I would not only consider running, but that I would consider running by the book. Rarer still, AROS’ roguelite elements and its full embrace of the positives of metagaming result in it being an adventure that benefits from repeated games, even with the same group. TL;DR: I recommend AROS to any dungeon-master whose players would enjoy experiencing their own, distinct form of dungeon-mastery.
Where to Find A Rasp of Sand
I am a new addition to the venerable Skeleton Crew. Here is a bit about me and what I am interested in exploring in my series (of which you have just finished the inaugural installment):
Pedantic Wasteland by WFS