A couple rules of thumb I use for crafting evocative descriptions as a GM:

THREE OF FIVE: Think about your five senses. Try to include three of them in each description. Sight is a gimme and a Taste will rarely apply, so that means picking a couple out of Hearing, Smell, and Touch. Remember that you don’t actually have to touch something in order to intuit what it might feel like if you did. Touch can also include things like wind and temperature.

TWO COOL DETAILS: Try to include two irrelevant-but-cool details. These are details that aren’t necessary for the encounter/room to function, but are still cool. It’s the broken cuckoo clock in the corner; the slightly noxious odor with no identifiable source; the graffiti scrawled on the wall; the bio-luminescent fungus; etc.

THREE-BY-THREE: Delta’s 1-2-(3)-Infinity talks about psychological research demonstrating that repeating something three times takes up the same space in our brains as repeating something infinitely. Thus, once you’ve hit the third item in a sequence, any additional items in that sequence are redundant.

Extrapolating from this, for minor scenes you can describe three things each with a single detail. At that point, you’ve filled up the “infinity queue” in your players’ brains and their imaginations, predicting the pattern, will impulsively fill in the finer details of the scene you’ve evoked. For “epic” descriptions, use the full three-by-three: Describe three different elements with three details each.

KEEP IT PITHY: On a similar note, lengthy descriptions are not how you achieve immersion. What you want are a handful of evocative images that the players can perform an act of closure upon to realize the scene in their own minds. This not only prevents them from “tuning out” a long description, the act of closure itself will draw them into and immerse them in the game world — they become, as Scott McCloud described under different circumstances in Understanding Comics, silent partners in the creation of the world, and thus become both invested in it and captured by it.

DESCRIPTION DOESN’T END: The legacy of boxed text indoctrinates a lot of GMs with the idea that you describe the environment up front and then, from that point forward, the players are just taking actions within the pre-established environment. But that’s an artificial dynamic. Instead, details of the setting can continue flowing through the description of that action: The crunch of broken glass beneath their feet as they move across the room. The coppery smell of blood as they draw near the corpse. The sudden chill from the wind whipping in through the shattered window.

Like most rules of thumb, of course, none of these should be treated like straitjackets.

This is an updated and expanded version of a tip that originally appeared in 2011.

I can’t do a murder mystery because the PCs will just cast speak with dead.

I’ve seen this sentiment a lot, but it’s never really made any sense to me: The act of investigating a mystery is one by which you reveal that which is unknown. When we talk about PCs casting a speak with dead spell, we’re describing a situation in which the players reveal that which is unknown (i.e., they investigate the mystery), but then, oddly, we’re supposed to conclude that they can’t investigate the mystery because they investigate the mystery.

I think part of the problem here lies in an erroneous instinct that I talked about as part of the Three Clue Rule:

There is a natural impulse when designing a mystery, I think, to hold back information. This is logical inclination: After all, a mystery is essentially defined by a lack of information. And there’s a difference between having lots of clues and having the murderer write his home address in blood on the wall.

But, in reality, while a mystery is seemingly defined by a lack of knowledge, the actual action of a mystery is not the withholding of knowledge but rather the discovery of knowledge.

Let me put it another way: Strip the magic out of this scenario. Imagine that you’ve designed a mystery scenario in which there was a witness to the crime. The PCs turn to this witness and say, “Who killed him?” and the witness says, “It was Bob.” And it turns out Bob is just standing there, so they arrest him. End of mystery.

You wouldn’t conclude that you can’t do mystery scenarios because people can talk to each other right?

Speak with dead should be no more alarming than an FBI team taking fingerprints or a CSI team enhancing video and running facial analysis.

DESIGN TO THE SPELL

You may also see people suggesting that you “nerf” the spell to one degree or another. (Corpses that refuse to answer questions, for example.) Nothing is more frustrating to a player than having their smart choices blocked because the GM has some preconceived notion of how they’re supposed to be investigating the crime.

But what you can do is design your mysteries to the reality of the spell. Generally speaking, after all, people in the game world know that the spell exists, right? So they aren’t going to plan their murders in ways that will expose them. (Any more than people in a magic-free setting will commit their murders while standing directly in front of surveillance of cameras.) They will find ways to conceal their identity; they may even find ways to try to use the spell to frame other people. (For example, imagine a murder scenario where the victim thinks one of the PCs did it because the perpetrator used a polymorph spell.)

OTHER DIVINATIONS

The same advice generally applies to other divination spells, too. The only divination effects which are truly problematic are those which allow you to contact omniscient beings and receive crystal clear information from them. Fortunately, these spells basically don’t exist in D&D (and most other games). The closest you can get are commune and contact other plane, but both are explicitly limited to the knowledge of the entity you’re contacting. (1st Edition AD&D actually had a lovely table for determining “Likelihood of Knowledge” and “Veracity”.)

FURTHER READING
Zone of Truth Mysteries
Three Clue Rule

Something I like to occasionally do during a session is to speak in tongues. It can be a nice touch of flavor to have one of the orcs the party is talking to turn to their comrade and whisper something in unintelligible orcish. Or to have an elf lord curse at them with silken invectives. Or have the strange, angelic being woken from an elder age quiz them in the stilted tones of a forgotten common tongue.

When I do this, of course, it’s not actually important that what I’m saying actually means anything. (If you actually know a fantasy tongue like Tolkien’s Quenya, that’s fantastic, but not required for this technique to be effective. And sprinkling in established words from a fantasy language as a sort of slang is a different thing, albeit also cool.) What is important is that the content flows, varies, and has a consistent tone. In other words, it needs to sound like someone actually speaking.

However, this can be difficult to smoothly achieve. To assist with this effect, I’ve created a tool I refer to as a Fantasy Lorem Ipsum: For each fantasy language, I have two pages of pre-generated text (which can be printed on both sides of a single sheet of paper). When I want to “speak” in that language, I can simply choose a location on the sheet and begin performing from it. You can also use these sheets to quickly generate handouts by copying and pasting a chunk of text. (You can then provide the “translation” separately if one of the PCs knows the language.)

Ancient Common
Dwarvish
Elvish
Halfling
Orc

The use of a GM screen can be a surprisingly contentious subject in the running of a roleplaying game. The critics consider them superfluous at best or intrinsically damaging to the dynamic of the game (due to inducing issues of trust and social separation) at worst. But I, personally, find them valuable more often than not, and I’d like to share my thoughts on how they can be used to best effect.

First, I don’t like the older style of portrait-oriented screens. Their height does, in my opinion, create an unnatural barrier between the GM and the players. They feel like a giant wall, cutting off the natural expression of body language.

Landscape-oriented screens, on the other hand, don’t have that problem. As the GM, I can see everything that’s happening on the table and the players can freely see my body language. As long as you’re playing with a table surface, there’s no meaningful disadvantage to the use of the screen and, in my experience, there are two significant advantages.

SIGHT LINE OBSTRUCTION

The most basic function of the screen is to block the player’s line of sight to my notes and maps. This is important to me not because I think my players are horrible cheaters who are trying to peek at my notes; it’s because I consider it a common courtesy. If I’m inviting people over to watch a movie, I don’t hang a poster with spoilers for the movie next to the TV screen and ask them to avert their eyes from it.

The same principle applies here. In fact, rather than inhibiting a personal connection between me and the players, I often find that a landscape screen enhances it: When you don’t obstruct your maps and such, players will often avert their eyes from your end of the table in order to avoid glancing at them.

VERTICAL REFERENCE

I’m a pretty huge advocate of being able to simultaneously display multiple pieces of information in order to facilitate rapid referencing and cross-referencing while running the game. (This is also why I don’t like running games from a laptop: The search functionality can be useful, but being able to only look at one page of information at a time while GMing is like trying to run a marathon with your legs tied behind your back.)

Therefore, being able to position reference material in a vertical place (so that it doesn’t take up surface space) is, in my opinion, insanely useful. In addition, positioning persistent reference material for the system and/or game world on the screen creates a consistent spatial familiarity that makes referencing that material faster and more efficient. (Instead of figuring out where the cheat sheet packet is currently lying on the table, picking it up, and flipping through it, I instead know that I can reach out to my right, flip up a piece of paper, and look directly at the skill difficulty guidelines. After just a couple of sessions, I basically don’t even have to think about it any more. It becomes autonomic.)

My typical table arrangement when GMing is:

• A customizable, landscape GM screen with four panels of information.
• 2-3 pieces of paper displayed behind the screen.
• One or more TV trays to my left side, which I use to hold my rulebooks and also display 4-6 additional sheets of information (which often includes one or more rulebooks flipped open to the appropriate page reference).

Without the GM screen, my quick reference material not only becomes less efficient, it also begins encroaching into the space I use for other reference material. This becomes a cascading problem, as useful resources get bumped out of circulation. With less information at my fingertips, it becomes more difficult to run complicated, interconnected scenarios.

MAKING THE SCREEN

As useful as the reference material on a GM screen can be, the sad reality is that most published GM screens feature a lot of non-essential information while not including material that would actually be useful when running the game. As a result, I use a modular, customizable landscape (like the ones you can buy here or here).

IMAGES: Buying the PDF version of an official landscape screen is often a good way to stock the player-facing side of your screen. But in an era of Google Image Search, the whole world of art and photography is your playground.

Personally, I tend to avoid trying to find single mural-style images that will go across the entire breadth of the screen. Finding multiple images to make up a polyptych is easier, and it also gives you the opportunity to highlight multiple facets of the game / world / campaign. I also recommend finding images that depict things the PCs could theoretically see, rather than images of main characters who aren’t the PCs doing awesome things. (It’s more immersive and suggestive of the game world that way, while allowing the table to remain focused on the narrative you’re creating instead of some other narrative that’s being depicted.)

(In the past I’ve also played with stocking the player-facing side of the screen with player-relevant reference material. But I’ve found that reading the material at any meaningful distance is usually difficult and, for players (who usually juggle less reference material), it’s easier to just use cheat sheet packets. Your mileage may vary.)

REFERENCE MATERIAL: I design System Cheat Sheets for many of the RPGs I run, particularly those featuring complicated mechanics. These reference sheets can then be conveniently slid into the modular screen.

A major conceptual breakthrough for me was the Hackmaster GameMaster’s Shield, which included flip-up panels:

Copying this same technique, I now use reverse-duplex printing to create sheets that I can tape together and flip up to reveal additional information behind them. This allows me to easily put 12 landscape-formatted sheets within easy reach. (And there’s no reason I couldn’t expand that to a third layer of information to give me 20 sheets, although I haven’t actually found a game so complicated that I would need to do that yet.)

FURTHER READING
Rolling Behind the Screen

One of the things I talked about in The Art of the Key is the conceptual organization of material from the general to the specific: What the PCs immediately see. What they might see. What they can investigate. What they see when they investigate X. This organization, in turn, naturally mimics the way in which the game space is explored during actual play. (Which means that the basic conceptual structure is useful whether you’re keying a location or improvising on the fly.)

I’ve recently realized that there is a specific elaboration upon this general structure that (a) I’ve been using in play for several years without really consciously thinking about it, and which (b) has proven to be very effective. I’m now referring to it as a Matryoshka Search.

Let’s say that PCs are exploring the bedroom of a serial killer. There’s a hidden trap door in the floor leading down to the killer’s mystic butchering chamber. The basic way of handling this scenario looks like this:

Player: There’s gotta be more here. I search the room.
GM: Give me a Search check.
Player: (rolls dice) 25
GM: You find a secret trap door in the floor.

That obviously works just fine. But what I’ll frequently end up doing instead is something like this:

Player: There’s gotta be more here. I search the room.
GM: Give me a Search check.
Player: (rolls dice) 25
GM: There are scuff marks on the floor around the legs of the bed.
Player: As if the bed had been moved back and forth a lot?
GM: Yeah.
Player: I shove the bed to one side and take a look.
GM: You find a secret trap door in the floor.

Instead of immediately discovering the item of interest, the character instead discovers an indicator pointing in the direction of the item of interest. The advantage is that it allows (and even requires) the player to receive information and then draw a conclusion. It’s a subtle distinction, but the result increases the player’s engagement and reduces the feeling that the GM is just handing them whatever information he feels like. I call it the Matryoshka search technique because it turns the interaction into a nested doll: One investigation “opens” new information, which can then be opened by another investigation in turn.

We could also look at this through the lens of the Art of Rulings: The GM is setting an initial threshold for player expertise activating character expertise which is fairly low. (All the player needs to do is say that they want to make a Search check.) But once the character’s expertise has given some sort of result, the GM hits the pause button, turns the interaction back to the player, and basically raises the required threshold. (“Your original declaration has taken you this far, but now I need more information.”) This can make the technique a good way of compromising between players who prefer a very low threshold of player expertise and GMs who want their players to engage more directly with the game world. It’s a naturalistic way of asking, “How are you doing that?” while still moving the action forward.

The Matryoshka technique works even when the indicator really only points at one possible conclusion, as it does in the example above. (Although even in the case of the bed being frequently moved there’s still the question of WHY the bed is moving.) But it can be even more effective if there are multiple explanations possible, requiring additional inquiry and thought before firm conclusions can be reached. As a very simple example, the GM might say, “Taking a closer look at the floor, you can see through the dust and grime clear indications of square-shaped seams.” Is it a pit trap? Is it a pedestal that rises up? Do the seams release poison gas or a force cage projected from below? The player is going to have to figure it out.

This technique is particularly valuable if you’re running the GUMSHOE system: Because every skill use in GUMSHOE is guaranteed to succeed, it can be very easy for investigation actions in the system to feel like “laundry lists” with the player simply naming the skills they want to use and then the GM handing them dollops of information. Matryoshka nesting of information can prevent the automatic successes from becoming lifeless.

FURTHER READING
Matryoshka Hexes