Tag Archives: technique

Types of Preparation

Download a draft PDF of my session document

What makes up the work that a referee does when preparing to run an RPG? Here is a taxonomy that I find useful for evaluating the utility of published products, and also for deciding how to spend my own limited time. I cover the categories from the general to the specific.


Atmosphere includes the kind of things that will define a setting at the highest level. Preparation at the this level is like basic scientific research. It is necessary if you don’t want stagnation, but is not very useful when the rubber meets the road in actual play. Luckily or unluckily, the vast majority of published RPG material is atmosphere. For example, most published tabletop RPG settings fit here. They are very far from being play-ready, though they might have some good ideas. Most game entities (the contents of “splatbooks”) also fit here, and include things like monsters, treasure, and spells. Even most modules are better situated here. Fun pleasure reading, interesting ideas, but often not so good at the table. I don’t want to denigrate atmosphere too much; you need to get your ideas from somewhere. But reading a module or game setting is often at the same level as watching a movie or reading a novel.


At some point, you need to start deciding where things exist in the campaign world. In the simplest case, you don’t need to do much work here; a town and a dungeon are enough for traditional D&D. Genre expectations (e.g.: generic Tolkienized medieval fantasy, Gotham City) can do much of the work for you, assuming that you don’t require your setting to be unique. The standard tradeoff here is approachability versus specialness (the same tradeoff exists for base rules and house rules). This level, for me, is no longer about general info (that would go above in the atmosphere category); the point of this is stuff that PCs might interact with at a macro level, both in spacial and relationship terms (e.g.: north of the kingdom are mountains, the guild of thieves seeks to steal the secrets from the council of magicians). There are few examples of published setting material under this definition. Most published “settings” are 90% atmosphere with 10% actionable setting info mixed within. I’m still not sure what the best way to store and reference setting info is, especially for use during the game. Character generation rules (or creation of pre-gens), selection of base system, and house rules all also fit here, practically speaking.

In an old school style game, this will likely be a site to explore, but it could also be something like an NPC relationship map. It must allow players to make low-level tactical decisions. In terms of published RPG material, the module is the most obvious analogue, but I’m coming to believe that the one page dungeon is a better model. Unfortunately, historically most published examples have been flawed by verbosity and linear story-based presentation that do not allow player choice to have much influence over how the game plays out. Verbose modules can still be valuable, but as atmosphere as described above.
This phase doesn’t have a published analogue that I have seen (pointers welcome!), and so it gets far less attention than it deserves. For many people (myself included, until relatively recently), this phase entails a few hastily scrawled notes, maybe a map, and perhaps some refresher cram-memorization if running a module. However, I find that I have been able to run games much more effectively given a slightly more structured approach. Specifically, I need to be able to track time and monster health. To assist with this, I roll up a set of encounters and hit dice beforehand (inspired by Jim’s DM prep posts over at Carjacked Seraphim and Courtney’s session tracker over at Hack & Slash). Turn sequence and hit dice are randomized before every session. This means that during play, I only need to check things off. It is surprisingly freeing to have this info predetermined, and I highly recommend it. Before I did this, I was unable to reliably track time. Afterwards, it became trivial. In addition to this tool, I sometimes create a list of more complicated encounters, compile a list of names to use for improvised NPCs, and have a section to note down treasure or “important things” discovered. This document is still a work in progress though, and I assume it will continue to evolve.

Monster Defaults

Parsimony is a virtue in monster stat lines. It seems like many game products make a fetish of following a template however, even if much is wasted space. For example, the omnipresent “Magic resistance: Nil” lines in 2E, though there are similar examples in all editions. This was recently brought up on G+, and I thought I would share how I do things here.

The basic idea is that I have a “default” monster (which is very close to a first level fighter), and I only specify anything that differs. Here is said monster:

# appearing 1, HD 1, AC unarmored, # attacks 1, damage 1d6, movement as unencumbered human (12), save as fighter of level equal to hit dice, attack as fighter of level equal to hit dice (or by monster hit dice, depending on the game), size as human (medium), % in lair, treasure none, morale 12 (fearless).

% in lair and size are actually not things that I have been specifying, but they were brought up and I think they are useful. I default to fearless for morale because that probably represents the single largest group of foes (undead, constructs, etc), even if they are not a majority (so most of my stat lines do end up with a different morale entry).

This leads to monster entries that look something like:

Robot, HD 3, AC as plate

And everything else is assumed.

I would probably include XP rewards too in anything intended for others, but I’ve been experimenting with so many different methods for rewarding XP that it would not be all that meaningful to me right now.

See also Alex Schroeder’s method.

Excursion Format

Back in high school, we sometimes played D&D in a format that we called a house game. This format was so called because every adventure was required to begin and end at a home base (the “house”). We did this because it allowed us to rotate DM duties, and slowly develop the campaign world jointly, rather than requiring a large time investment upfront by a single DM. This was the way we ran the Blackwater Falls campaign.

This is somewhat similar to the way I see many games being run online now (ConstantCon, FLAILSNAILS, etc), though the primary concern is not rotating referees, but rather a changing player roster. The PCs in the next session may not be the same as the PCs in the previous session. In other words, it is assumed that PCs return to town (whatever “town” means for the particular campaign).

This can raise a problem of logical narrative. What happens if the PCs end the session on, for example, dungeon level 3? Jeff Rients addresses this with his Triple Secret Random Dungeon Fate Chart of Very Probable Doom. Basically, if you are still in the dungeon at the end of the session, you need to make a roll (50% success) adjusted down (-10%) by dungeon level and up (+10%) by character level to see if you escape without mishap. If you fail that roll, you are sent to the Chart of Very Probable Doom to see what happens (the name says it all).

Jeff’s approach and table are inspired, but I am modifying the method slightly for my own use. First, rather than making a percentile roll and adjusting for level, the character makes a saving throw (the most favorable save may be used). This models the increasing competency of being higher level and reuses the numbers that are already on the character sheet (not that Jeff’s percentages are hard to remember; I just like saving throws). You may also add your single most favorable ability modifier as a bonus. Penalty is applied by dungeon level as per Jeff, but only -1 (5%) per level (I may change this to -2 in the future). Second, regarding the wilderness. The number of hexes away from the nearest known point of civilization is equivalent to dungeon level in terms of penalty. Third, the expected consequence of failure is death. I may have a table with other options, but its contents are confidential.

Note that whatever system I happen to be running, it is likely that the main setting assumptions of B/X D&D will hold: the wilderness is a perilous place. You might run into a dragon. I stock the wilderness map without consideration of PC levels. Merchants and other travelers generally move in large, armed caravans. Maybe you should stick to the dungeon for the first few levels (not that that’s any guarantee of safety either).

This post is not intended to be in any way original. I just want a page that I can reference describing the way I plan to run a game.

9 July 2012 edit: Well, look at this, an ODD74 thread from 2008 (Starting my Jakalla Megadungeon) that features a “table of despair” for characters that have overstayed their welcome in the dungeon.


I have found that I don’t much like rolling on content-generating tables during play. I rarely do it (I would usually rather just improvise), but every once in a while it comes up (like a treasure table in a module that I didn’t bother to roll on beforehand). The box label generator in the Lamentations module Tower of the Stargazer is a good example of this. Rolling those names during the game just killed the sense of immersion, and made it seem like none of the results could possible matter.

Whenever this happens, I feel like it slows down the game and exposes pieces of the machinery that are better left hidden. My most recent face to face group, especially, seemed to become uninterested in content if it looked like the random variety. They wanted to find the “real” content.

There have been a few posts on the blogs recently that have touched on similar issues. For example, Beedo over at Dreams in the Lich House has been talking about how he is using spreadsheets to pregenerate content for his Black City game:

And, Jim at Carjacked Seraphim has been posting about his system for DM prep. He has some useful-looking ideas there, like prerolling on which turn encounters should happen, so that you can tick off bubbles as the turns progress and then just cross-reference the appropriate encounter column. Check them out:

This also brings an aspect of fate into the game without actually limiting player choice at all, which is sort of fascinating. It’s like looking down from the corner of a high building at a road intersection and seeing two cars speeding towards (but oblivious of) each other. You have seen their future (the crash) without reducing their free will.

There are also tools like Meatshields! that can help.

The principle is also a bit like vancian casting: you want to prepare the content so that all that is left is the final command word. Note that the content in question can still be loosely bound. Like, you might not know exactly where you are going to need the next barkeep, but having one ready is useful (especially if you are as bad at remembering improvised details like I am).

Constraint & Creativity

As anyone who has engaged in creative endeavors probably knows, boundless freedom is often not an aid to creativity. Instead, limits and strictures seem to help channel ideas from chaos into some semblance of meaning and potential newness. Paradoxically, censorship is even a form of constraint which can foster creativity (especially clever ways of communicating that which is prohibited). This expands on my previous post about persistent settings, where I touched on the idea of constraint briefly.

I think constraints function in two main ways to help facilitate creativity. The first is that constraints often give you a place to start, helping to bypass the blank sheet problem. The second is that the task at hand is narrowed down to reconciliation of desired effect with particular limits. These properties should be familiar to people who have studied productivity techniques; methods to get started (getting past the blank page) and methods to break larger, complex tasks into smaller, simpler tasks.

Mixing results from random tables is thus a method of introducing constraints. How are these disparate results reconciled? How does it make sense that there are berserkers in the first room and goblins in the second room? Why is a dragon encountered only six miles from a town? Is it perhaps the hidden servant (or master) of the town mayor? Why is there a desert right next to the sea? How does the isolated town support itself? Matt Finch calls this process deep design in his Tome of Adventure Design (one of my favorite RPG books; I have not spent nearly enough time with it).

Some other ideas for limitations:

  1. Limit yourself to a core rulebook or boxed set. I’m leaning towards using the OD&D 3 LBBs (I already have a basic alignment-based taxonomy to use as an organizing principle).
  2. Only take monsters from one (non-standard) bestiary (there has been some blog discussion about this over the past few months regarding the Fiend Folio).
  3. Only use certain tools during creation. Scott Driver is doing this with his Dwarf-Land setting by using a typewriter. One could also hand-write everything.

Mapping to the Battlemat

As you probably know, miniatures and a grid are generally assumed by Fourth Edition. They are not strictly speaking required for playing a 4E game (it is possible to run 4E combat entirely using imagination), but my players seem to like using the battle mat. Using miniatures is relatively new for me, as we never used minis back in the 90s when I played Second Edition. Everything was imagination and description, with the occasional sketch for clarification.

I currently use a Paizo GameMastery Flip-Mat. This is a dry-erase battle mat with dimensions of 24 x 30 inches. Now that I think about it, it seems like I would save myself some time if I used these same dimensions on my one page dungeons. I suppose this should be one of those self-evident things, but took me 9 months to realize (I my defense I’ve also been running lots of converted modules written for other systems). Defaulting to this size doesn’t restrict the overall size much, though it does place some constraints on individual rooms and encounter areas, as 5 foot squares results in 120 x 150 feet. This is really not that large of an area.

One danger of mapping to the mat is that players might figure out that maps tend to have these dimensions, and thus engage in metagame reasoning (“we should turn left here because that side of the mat is unexplored”). While I don’t consider metagame reasoning to be inherently bad, I do think it can take away from immersion in some cases, especially if it is happening during play (as opposed to deciding which feat to take or something like that). To combat this, one should periodically make partial battle mat maps. Keeping the overall dimensions in mind is still useful though, even in this case.

I have had two other ideas recently regarding handling the battle mat and miniatures in play. The first idea is to delegate the mat drawing duty to a player rather than doing it myself. I think this might speed things up and also increase player engagement. They would need to create the tactical map from my verbal description, though I could of course correct obvious inaccuracies. This also reminds me of how James from Grognardia has his players assist with creating models of dungeon areas while he is engaged in verbal description.

The second idea is switching to a gridless battle mat. I think there is good value in being able to see spacial relationships. What I’m less sold on is the numerical calculation that comes with counting movement squares and areas. I feel like this is the part of grid play that can potentially hurt immersion and game flow. It allows a sense of certainty that should not be present in a combat situation. For example, if you know the enemy has a move of 6 and you are 7 squares away. Now, one could always break the rules and vary NPC movement rates (or really anything) but I don’t like doing that. I’d rather have a bit of uncertainly built into the basic experience, and I think using a gridless mat might help with that (using common sense for things like movement distances and effect areas).

Play Aids

Following on my ideas for one page dungeon module prep, I’ve actually put together some materials. That’s 10 pages in the picture below (I printed the Blasphemous Brewery prep sheet double sided because I was able to fit the maps and prompts needed on two sheets). One of those sheets is a table for the effects of the purple lotus. I’m missing zone 2 of Hammers of the God because I haven’t completed it yet. It might need to be two pages; I’m not sure yet. The transfomorph from that module also has its own page because it requires several tables.

I’m in the process of running Death Frost Doom from these sheets. They are working out really well so far. They do take a bit of effort to construct, but if you do it at the same time you are reading the module in preparation, I don’t think the extra time is all that much. I would love to see this sort of play aid generally included with modules. Basically, I think that page flipping should be kept to an absolute minimum. It is a rare case where a session might require more than one or two of these sheets (if players are in a sandbox environment where they could jump in any number of different directions, you might need to have several available, but you will probably not need to actually use more than one or two).

As I said before, I don’t think these are replacements for traditional adventures. Modules, especially good ones, convey a lot of atmosphere with all that extra text. But I don’t want to read that during play, I want to read that before play to have a sense of what the scenario should feel like, and during play I only want the critical details.

I’m still playing around with the form factor. Assuming a single sheet like this works well, but perhaps digest size, printed like a booklet (one sheet of paper folded in half) would work better. That might make the front and back more accessible, and would also allow the sheet to be kept inside my adventure log notebook, which is smaller than 8.5 x 11. Right now I have a binder in addition to the notebook, and it would be nice to only need the notebook.


Older stories in the fantastic tradition often feature a protagonist that is not of the fantasy world. Alice in Wonderland. A Connecticut Yankee in King Arthur’s Court. The children that travel to Narnia. John Carter of Mars. Occasionally, a more recent story will also employ this technique (for example, Terisa in The Mirror of Her Dreams), but in general this method is rarely used now. I believe this development is tied to the idea that fantasy settings should be complete within themselves and internally consistent; a replacement for reality rather than an interesting location meant to function as a stage.

Originally, the technique of taking a “real world” person and inserting them into a fantastic setting was probably meant as a way to help the audience accept the fantastic world by giving them someone familiar (the protagonist) to identify with. This structure also works well because as a visitor, the protagonist is not expected to already be familiar with the setting, so they can discover the setting along with the reader. John Carter didn’t get a setting background book when he was transported to Barsoom. This is a good fit for tabletop RPGs: no setting infodump assimilation required.

John Carter and Alice are both essentially planar travelers. This is also more or less what FLAILSNAILS characters are. I used to worry about making sure that all PCs “made sense” in the setting, but I’m coming to care less and less about this. The FLAILSNAILS conventions have taught me that my players can run whatever they want without messing up my campaign setting. My current players have been making characters using the online 4E character builder, which offers a huge variety of races, classes, and powers. None of them have made anything too exotic yet, but with this framework it wouldn’t even matter if they did. Just treat any odd PC the same way you would treat a FLAILSNAILS ConstantCon character.

OPD Modules

Back in my 2E days, I used to almost never run modules. The only module I can remember using was this Ravenloft module RA3 Touch of Death. But I’ve been using modules more recently, and the experience really is quite different than writing your own scenarios. My current 4E game was my first experience as a referee using the system, so to begin with I wanted to use a module to get a sense of how the system was expected to be used. I ran part of Seekers of the Ashen Crown (really, I just borrowed the first dungeon and discarded the plot railroad). This was before I had discovered the OSR, and I was just getting my feet wet after having not played for upwards of 10 years.

Since then, I have run Pod-Caverns of the Sinister Shroom and Tower of the Stargazer (a B/X one-shot unconnected to my current campaign). Both of these were very successful and I feel like I’m getting the hang of making a module my own, but the utility of the traditional module format is not very good. As artifacts to use at the game table (either PDF or hard copy), different kinds of information are jumbled together and page flipping is constantly required. Many new school players talk about innovation in game design and how we should not be slaves to nostalgia. Like in computer engineering, they say, sometimes you need to break old interfaces (and discard compatibility) in the service of progress. When I think of actual progress in game design though, the first thing that comes to mind is the one page dungeon (OPD) format.

I wrote before about some techniques for module preparation. Basically, boiled down, that post amounts to annotating modules by separating the different types of information. During play, one should not need to digest large blocks of text, as that slows down the flow of the game. My recommendations in that past post do help, and the process of summarizing (much like taking notes when reading) also helps to assimilate information. It is a way of reading actively rather than reading passively. For those of us without photographic memory, this can make a big difference regarding retained knowledge.

We can still do better though. For my next adventures, I am planning to translate the modules into the OPD format for use during play. Some might object that the work required is almost as much work as creating a scenario from scratch, but I disagree. Such a position undervalues the underlying creativity required to create a really engaging scenario. Of course, I haven’t done such a translation yet, so we’ll see how much work it takes. Of the published modules I have read so far, I feel like Stonehell Dungeon comes the closest to my ideal format (it uses a two page facing variation on the one page dungeon) though the OPD pages don’t have many memory cues for descriptive elements.

There is a place for detailed textual descriptions like are standard in traditional modules, but I feel like the optimal referee play-aid should almost never require a context switch (i.e., page turn). So, for me, a perfect module would be a textual overview followed by detailed descriptions of areas and NPCs. In addition, there would be an OPD for every zone which would include short-form stats and basic reminders about details. A facing page (or on the reverse) could include extra details for encounters with more moving parts (like a potion rack with lots of different possible effects or a puzzle). In addition, there should be a map-only version of the zone OPD to facilitate restocking. The best tool for getting the feel of a location is not the best tool for actually running that location. Think about the difference between a novel and a script.

Tracking Resources

According to Gary Gygax (Dungeon Masters Guide page 37, caps in the original):

One of the things stressed in the original game of D&D was the importance of recording game time with respect to each and every player character in a campaign. In AD&D it is emphasized even more: YOU CAN NOT HAVE A MEANINGFUL CAMPAIGN IF STRICT TIME RECORDS ARE NOT KEPT.

Tracking resources is hard, at least for me. Once events start getting complicated, resource tracking is generally the first thing to be jettisoned. It seems that other people have similar problems.

When it comes down to it, there are only a few core resources:

  • Time
  • Food
  • Water
  • Ammunition (arrows, quarrels, daggers)
  • Light (torches, lantern oil)
So here’s an idea. Use a stack of poker chips for each resource (regarding poker chips, see also: Lord KilgoreTelecanter, Lord Kilgore again). That way, there is a visual cue for the decreasing resource (as the stack gets shorter). Optimally, each character would have a stack for each resource, with the “time resource” belonging to the referee. One could also use other kinds of tokens or counters, but I like the visual representation of a stack (height is a very powerful metaphor that affects our thinking in many ways).

Proposed poker chip equivalences:

  • Yellow = oil (1 hour of light, 4 chips per flask)
  • Red = torch (1 hour of light, 1 chip per torch)
  • Blue = drinking water for 1 day
  • Green = rations for 1 day
  • White = ammunition (1 per arrow)
This could be modified in various ways depending on the precision desired. For example, for exact tracking, light units could be measured in turns. Thus, each torch carried would be represented by six red chips. A lit torch would result every turn in a red chip being transferred from the player’s torch stack to the referee’s time stack. To see how much time has passed, one need only look at the time stack. I’m not sure how this would work in play. It might be annoying to move a chip for every turn (on the other hand, it might make time more salient).
Exact tracking would require 24 chips per flask of oil though, which is probably too many (though one could have a stack representing flasks and another stack representing the active lantern). It might be more reasonable to store light in hour units (1 per torch, 4 per flask of oil). The referee would need a separate way of ticking off turns that pass behind the screen, but my guess is that such a method would work better. One could model time using two stacks: one for days and one for hours. Once the appropriate number of hour chips have accumulated, the players know it is time for the PCs to rest (or push on with exhaustion penalties). Alternatively, one could have a pile for the active torches or lanterns with one chip per turn of light and a separate pile for spare torches or flasks of oil. This also has the advantage of the players being able to watch their torches burn down turn by turn.

I really need to try this to see how convenient it is. My players are still down in The Pod-Caverns of the Sinister Shroom, and I don’t want to introduce such a mechanic in the middle of a delve (I’m pretty sure none of them have rations written on their character sheets, for example, which I am magnanimously overlooking).
Tracking ammunition with this system seems straightforward. Secondary ammunition (like a PC that carries a crossbow and multiple throwing knives) could be represented by other chip colors. If a PC carries more than 20 shots of a particular kind of ammunition, just track the active quiver (like the active lantern example above). And really, unless you have a wagon or a retainer, I think it’s highly unlikely that one would carry more than 40 (or even 20) arrows, especially when carrying other gear. Arrows are bulky.
It seems like this system would work just as well for wilderness journeys, as the primary resources required for overland travel are food and water, which are usually measured in days. Passing days are then represented by every player decrementing their water and food piles by one, and the referee incrementing the days pile.

What I think is interesting about this approach is how it illustrates the action of a ledger: spent resources (light, food) are transformed into passing time. Various abilities can also more easily “cost time” using this approach. Traditionally, actions like searching for secret doors are supposed to cost time. For more on time as a resource, see JB’s post on the “automatic” thief.

Am I missing any resources that are important to track?